A high-risk weather identification method and device based on monitoring data of a power transmission line micro weather station

Abstract

The application discloses a high-risk weather identification method and device based on power transmission line micro-weather station monitoring data, which comprises the following steps: collecting power transmission line micro-weather station monitoring data, and segmenting the data into multiple data segments through time sequence segmentation to extract weather events in each data segment; according to a high-risk weather event marking strategy, marking the weather events in the extracted data segments to obtain a part of data segments containing high-risk weather events; extracting the description features of all data segments, analyzing and processing to form a core feature set; combining the core feature set and the part of data segments containing high-risk weather events to perform cluster analysis on the high-risk weather events; and obtaining the characteristic parameters of the high-risk weather events in time and space to complete the high-risk weather identification based on the power transmission line micro-weather station monitoring data. The application can evaluate the influence of weather events on the power system in a small scale dimension, effectively guide disaster warning and deployment processing, and has a wide application range.

Description

Technical Field

[0001] This invention belongs to the field of power equipment monitoring and early warning technology, specifically relating to a method and device for identifying high-risk weather based on monitoring data from micro-meteorological stations along power transmission lines. Background Technology

[0002] Transmission towers, as the most important infrastructure in power transmission networks, bear the responsibility of supporting conductors for power transmission and constructing the power grid. Towers are typical wind-sensitive structures, and their damage is often related to extreme wind conditions. Regarding high-risk meteorological analysis, the power sector currently relies primarily on publicly available meteorological information, supplemented by meteorological monitoring stations, for severe storm analysis.

[0003] For example, patent CN114626629A discloses a method for assessing the risk of meteorological disaster faults in transmission lines. It establishes an analytic hierarchy process (AHP) assessment model based on the influencing factors of meteorological disasters, calculating the risk weights of these factors. Then, it uses the PageRank algorithm to establish an algorithmic assessment model based on these factors, calculating the corresponding risk consequences. Finally, based on the risk weights and the risk consequences, it establishes a comprehensive operational risk assessment model for transmission lines. This method improves the accuracy of meteorological disaster risk assessment for transmission lines, enabling accurate understanding of the risk status of the lines and timely assessment of the safety level of line operation under multiple meteorological disasters.

[0004] For example, patent CN114330061A discloses a method for analyzing the weak points of typical transmission line towers. This method involves acquiring geographical environmental data and establishing a corresponding geographical environment model; acquiring data on transmission line towers and erected cables and establishing corresponding transmission line tower and cable models; acquiring meteorological data corresponding to the geographical location and establishing a corresponding meteorological model; and analyzing and judging the weak points of the transmission line towers based on the influence of the meteorological model on the transmission line tower and cable models, and the combined influence of the meteorological model and the geographical environment model on the transmission line tower and cable models. This solution proposes a multi-factor analysis method for different meteorological conditions in different regions, suitable for finding and analyzing the mechanical weak points of typical transmission line towers, effectively identifying weak points in transmission line towers, and improving the safety of transmission line tower operation.

[0005] However, the existing meteorological technologies mentioned above identify strong winds at the mesoscale level and study severe storm warning methods based on mesoscale meteorological numerical forecasts. These warnings only remain at the mesoscale level and cannot assess the impact of meteorological events on the power system at a smaller scale, thus failing to effectively guide pre-disaster warning and post-disaster resource allocation. Wind-to-tower response data collected from power towers using mechanical sensors suffers from high costs, complex designs, and poor stability in data acquisition equipment. Furthermore, the research methods are highly correlated with the tower model and structure, making widespread application difficult.

[0006] Transmission line micro-weather stations are a meteorological monitoring technology that utilizes transmission towers to collect data on wind speed, precipitation, and temperature, enabling real-time monitoring of weather conditions at a height of 10 meters above the ground. Currently, some researchers are beginning to explore and analyze data from transmission line micro-weather stations to provide decision support for the planning and operation of overhead transmission lines in power systems, meteorological modeling, and post-disaster analysis.

[0007] For example, patent CN113408788A discloses a method, system, device, and medium for high-dimensional construction and completion of micro-meteorological monitoring devices. The method includes the following steps: acquiring the geographical latitude and longitude information of the micro-meteorological monitoring device locations and analyzing the regional and temporal characteristics of micro-meteorology; performing cluster analysis based on two-dimensional latitude and longitude information on the micro-meteorological monitoring devices according to the geographical latitude and longitude information, classifying micro-meteorological monitoring devices with similar spatial geographical distances into the same category; acquiring the monitoring information of the micro-meteorological monitoring devices classified into the same category; constructing a three-dimensional missing tensor based on the monitoring information; and filling in the missing values ​​of the micro-meteorological monitoring devices based on the three-dimensional missing tensor and low-rank tensor completion algorithm. This invention, based on the spatiotemporal correlation of micro-meteorological monitoring information, fills in missing values ​​of erroneous micro-meteorological monitoring information and can be widely applied in the field of meteorological disaster early warning technology.

[0008] However, existing technologies have not yet studied how to process micro-meteorological data to accurately identify high-risk weather conditions for power transmission towers.

[0009] Therefore, how to design a high-risk weather identification method based on transmission line micro-weather station monitoring data, and associate the transmission line micro-weather station monitoring data with high-risk weather, so as to assess the impact of meteorological events on the power system at a small scale, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0010] To address the shortcomings of the existing technologies, this invention provides a method and device for identifying high-risk meteorological events based on monitoring data from micro-meteorological stations along power transmission lines. The method extracts meteorological events through time-series segmentation, labels meteorological data based on meteorological disaster information, extracts features of high-risk meteorological events, and uses cluster analysis to detect historical high-risk meteorological events that may damage power transmission towers. It assesses the impact of meteorological events on the power system at a small scale, effectively guiding disaster early warning and dispatching, and has a wide range of applications.

[0011] In a first aspect, the present invention provides a method for identifying high-risk weather conditions based on monitoring data from micro-weather stations along power transmission lines, comprising the following steps:

[0012] Collect monitoring data from micro-meteorological stations along power transmission lines, and divide the data into multiple data segments by time series segmentation, then extract meteorological events from each data segment;

[0013] Based on the high-risk meteorological event labeling strategy, meteorological events in the extracted data segments are labeled to obtain a portion of the data segments containing high-risk meteorological events;

[0014] Descriptive features are extracted from all data segments, and analyzed to form a core feature set;

[0015] By combining the core feature set and a portion of the data segments containing high-risk meteorological events that were labeled, cluster analysis was performed on each high-risk meteorological event;

[0016] Acquire the temporal and spatial characteristic parameters of high-risk meteorological events to complete the identification of high-risk meteorological events based on monitoring data from micro-meteorological stations along power transmission lines.

[0017] Furthermore, the data is segmented into multiple data segments by time series data segmentation, and meteorological events are extracted from each data segment. The specific steps include the following:

[0018] The monitoring data from the micro-meteorological stations along the transmission lines for each data segment are set to follow a normal distribution;

[0019] According to the time series, the monitoring data of micro-meteorological stations on transmission lines are searched for breakpoints, and the final breakpoint set is determined by the maximum threshold of probability likelihood estimation.

[0020] Based on the breakpoint set, the monitoring data of the micro meteorological station of the transmission line is divided into multiple data segments, and the corresponding meteorological events are identified and extracted in each data segment;

[0021] The process involves searching for breakpoints in the micro-meteorological station monitoring data of transmission lines according to the time series sequence, and determining the final breakpoint set by using the maximum threshold of probability likelihood estimation. The specific steps are as follows:

[0022] By using the time series sequence, a combination of breakpoints composed of each initial breakpoint is randomly generated;

[0023] By performing global and local searches on the monitoring data to update the breakpoints, the final set of breakpoints is determined.

[0024] The global search generates candidate breakpoints, while the local search performs probability likelihood estimation on each candidate breakpoint to determine the final breakpoint.

[0025] Furthermore, high-risk weather event labeling strategies specifically include:

[0026] Obtain the time and latitude / longitude data of the actual disaster event, compare it with the meteorological events in the data segment, and determine the corresponding data segment as the data segment containing high-risk meteorological events.

[0027] Furthermore, the descriptive features include static time-frequency domain features and dynamic distortion similarity features.

[0028] Furthermore, descriptive features are extracted from all data segments, and the data is analyzed and processed to form a core feature set. This process includes the following steps:

[0029] The time-frequency domain characteristics of meteorological events in the extracted data segments are analyzed to obtain the indicator feature set;

[0030] Based on the dynamic distortion similarity of meteorological events in pairwise data segments, a set of morphological features is obtained;

[0031] The indicator feature set and the morphological feature set are merged to form an initial feature set, which gives the similarity of each meteorological event;

[0032] All meteorological events were clustered using cluster analysis, and pseudo-labels were obtained for each meteorological event.

[0033] Based on the pseudo-labels, a random forest is used to filter the initial feature set to obtain the core feature set of meteorological events.

[0034] Furthermore, the time-frequency domain characteristics of meteorological events in the extracted data segments are analyzed to obtain an indicator feature set. This also includes dimensionless processing of the indicator feature set data, using the following formula:

[0035] ;

[0036] in, The raw data for the indicator features, for Dimensionless data of indicator characteristics for The mean of the indicator characteristics, for Standard deviation of indicator characteristics.

[0037] Furthermore, the indicator feature set includes a time-domain feature set and a frequency-domain feature set. The time-domain feature set includes waveform indicators, peak indicators, impulse indicators, margin indicators, skewness indicators, and kurtosis indicators. The frequency-domain feature set includes centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation.

[0038] Furthermore, the indicator feature set and the morphological feature set are merged to form an initial feature set, which gives the similarity of each meteorological event. Specifically, the similarity of each meteorological event is measured by the distance between data segments. The distance relationship between data segments is as follows:

[0039] ;

[0040] in, This represents the distance between two data segments in the time-frequency domain feature space. The cost of morphological transformation between the two data segments. Let n be the number of data segments.

[0041] Furthermore, combining the core feature set and a portion of the labeled data segments containing high-risk meteorological events, cluster analysis is performed on each high-risk meteorological event, specifically including the following steps:

[0042] Provide the core feature set of a portion of the labeled data segments containing high-risk meteorological events, and obtain the core feature set of high-risk meteorological events;

[0043] The similarity of all core datasets with the core dataset of high-risk meteorological events is compared, and the clustering results of other high-risk meteorological events in the core datasets are given.

[0044] Secondly, the present invention also provides a high-risk weather identification device based on monitoring data from micro-meteorological stations along transmission lines, employing the high-risk weather identification method based on monitoring data from micro-meteorological stations along transmission lines as described above, including:

[0045] The data acquisition module collects monitoring data from micro-meteorological stations along power transmission lines and segments the data into multiple data segments using time-series data segmentation.

[0046] The annotation module extracts meteorological events from each data segment, and annotates the extracted meteorological events in the data segments according to the high-risk meteorological event marking strategy, thereby obtaining a portion of the data segments containing high-risk meteorological events.

[0047] The analysis and processing module extracts descriptive features from all data segments, analyzes and processes them to form a core feature set, and combines the core feature set with a number of labeled data segments containing high-risk meteorological events to perform cluster analysis on each high-risk meteorological event, obtain the temporal and spatial feature parameters of the high-risk meteorological events, and complete the identification of high-risk meteorological events based on the monitoring data of micro meteorological stations on power transmission lines.

[0048] The present invention provides a method and apparatus for identifying high-risk weather based on monitoring data from micro-meteorological stations along power transmission lines, which has at least the following beneficial effects:

[0049] (1) Meteorological events are extracted by segmenting time series, meteorological data are labeled according to meteorological disaster information, and then the characteristics of high-risk meteorological events are extracted. Cluster analysis method is used to detect high-risk meteorological events that may cause damage to power transmission towers in the past. The impact of meteorological events on the power system is assessed at a small scale, which can effectively guide disaster early warning and dispatching. It has a wide range of applications.

[0050] (2) Cluster analysis is performed by the distance and morphological similarity of data segments in the time-frequency domain feature space, and the prediction of high-risk meteorological events in each data segment is given based on this. The accuracy is high and the data is highly operable.

[0051] (3) By sorting and analyzing the data from the micro-meteorological stations of the transmission line, and combining it with meteorological events, the extraction and detection of high-risk meteorological events can achieve accurate positioning from the two dimensions of time and latitude and longitude. Attached Figure Description

[0052] Figure 1 A flowchart illustrating a high-risk weather identification method based on monitoring data from micro-meteorological stations along power transmission lines, provided by this invention.

[0053] Figure 2 A schematic diagram of the process for extracting meteorological events from data segments according to one embodiment of the present invention;

[0054] Figure 3 A graph showing the results of extracting curves containing high-risk meteorological events from a data segment provided in one embodiment of the present invention;

[0055] Figure 4 A curve result graph containing no high-risk meteorological events is extracted from a data segment provided in one embodiment of the present invention;

[0056] Figure 5 A schematic diagram of the data segment description feature extraction process provided in one embodiment of the present invention;

[0057] Figure 6 A graph showing the results of cluster analysis on a data segment provided in a certain embodiment of the present invention.

[0058] Figure 7 A temporal distribution diagram of high-risk meteorological events provided in one embodiment of the present invention;

[0059] Figure 8 This is a schematic diagram of a high-risk weather identification device based on monitoring data from a micro-meteorological station on a power transmission line, provided by the present invention. Detailed Implementation

[0060] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0061] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0062] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0063] By linking monitoring data from micro-meteorological stations along power transmission lines with high-risk meteorological events, the impact of meteorological events on the power system can be assessed at a small scale. This requires analyzing the micro-meteorological monitoring data from these stations, extracting high-risk meteorological events that could cause damage to the power transmission system based on information about wind-induced tower collapses, and then identifying and statistically analyzing high-risk meteorological events during the operation of the micro-meteorological stations along the power transmission lines.

[0064] like Figure 1 As shown, this invention provides a method for identifying high-risk weather conditions based on monitoring data from micro-meteorological stations along power transmission lines, comprising the following steps:

[0065] Collect monitoring data from micro-meteorological stations along power transmission lines, and divide the data into multiple data segments by time series segmentation, then extract meteorological events from each data segment;

[0066] Based on the high-risk meteorological event labeling strategy, meteorological events in the extracted data segments are labeled to obtain a portion of the data segments containing high-risk meteorological events;

[0067] Descriptive features are extracted from all data segments, and analyzed to form a core feature set;

[0068] By combining the core feature set and a portion of the data segments containing high-risk meteorological events that were labeled, cluster analysis was performed on each high-risk meteorological event;

[0069] Acquire the temporal and spatial characteristic parameters of high-risk meteorological events to complete the identification of high-risk meteorological events based on monitoring data from micro-meteorological stations along power transmission lines.

[0070] Based on the statistical results of this invention, targeted operation and maintenance guidance can be provided for transmission towers, guiding relevant departments to rationally allocate resources and promptly eliminate potential risks to transmission towers. Thanks to the universality of the analysis object, this invention can be promoted for various types of transmission towers in different regions.

[0071] Meteorological event extraction uses time series segmentation to divide the monitoring data of micro-weather stations on transmission lines into smaller, interpretable data segments. Each data segment corresponds to a meteorological event and has a relatively consistent data pattern.

[0072] like Figure 2 As shown, multiple data segments are formed by dividing the time series data, and meteorological events are extracted from each data segment. The specific steps include the following:

[0073] The monitoring data from the micro-meteorological stations along the transmission lines for each data segment are set to follow a normal distribution;

[0074] According to the time series, the monitoring data of micro-meteorological stations on transmission lines are searched for breakpoints, and the final breakpoint set is determined by the maximum threshold of probability likelihood estimation.

[0075] Based on the breakpoint set, the monitoring data of the micro-weather station on the transmission line is divided into multiple data segments, and the corresponding meteorological events are identified and extracted in each data segment.

[0076] The process involves searching for breakpoints in the micro-meteorological station monitoring data of transmission lines according to the time series sequence, and determining the final breakpoint set by using the maximum threshold of probability likelihood estimation. The specific steps are as follows:

[0077] By using the time series sequence, a combination of breakpoints composed of each initial breakpoint is randomly generated;

[0078] By performing global and local searches on the monitoring data to update the breakpoints, the final set of breakpoints is determined.

[0079] The global search generates candidate breakpoints, while the local search performs probability likelihood estimation on each candidate breakpoint to determine the final breakpoint.

[0080] The maximum threshold can be preset before making a judgment. The preset value can be selected according to different application scenarios, and no further restrictions are made here.

[0081] like Figure 3 As shown, the third segment corresponds to the data of the meteorological event that caused the power transmission tower to collapse. In this segment, the wind speed rises rapidly in a short period of time, reaches its maximum value and causes damage to the power transmission tower, and then drops rapidly. This data characteristic corresponds to the process of a strong convective cyclone passing through, causing damage, and then moving away. The other segments correspond to other meteorological events that occurred successively during this period, and each data segment also corresponds to the entire process of a meteorological event from its inception to its development and eventual conclusion.

[0082] like Figure 4 As shown, no severe meteorological events occurred near the micro-meteorological station of the transmission line during the corresponding dates. The segmentation results of the time series can not only completely identify meteorological events with long durations, but also accurately detect meteorological events with shorter durations in the fifth segment. Moreover, each data segment corresponds to the occurrence process of meteorological events in terms of wind speed, from low to high and then from high to low.

[0083] High-risk weather event labeling strategies include:

[0084] Obtain the time and latitude / longitude data of the actual disaster event, compare it with the meteorological events in the data segment, and determine the corresponding data segment as the data segment containing high-risk meteorological events.

[0085] A power transmission tower collapsed near the corresponding data collection time period due to severe convective weather. The transmission line micro-meteorological station collected corresponding data fluctuations at that time, extracted the data, and marked it as a high-risk meteorological event waveform. After extracting the meteorological event, it is necessary to label it as a high-risk meteorological event based on the actual disaster. The disaster events originated from wind-induced tower collapses reported by various local departments of the power system; the disaster severity was high, and the corresponding meteorological events were highly hazardous. High-risk meteorological events at the actual disaster location often correspond to tornadoes, strong convective cells, and strong tropical typhoons. The wind speed range at the moment of collapse is often too small, making it difficult for the maximum wind speed to be directly captured by the transmission line micro-meteorological station. However, the wind speed waveforms of nearby transmission line micro-meteorological stations often have similar characteristics. In this embodiment, based on the information from transmission line micro-meteorological stations around the disaster location, the relevant meteorological events are labeled as high-risk meteorological events.

[0086] The descriptive features include static time-frequency domain features and dynamic distortion similarity features. Static time-frequency domain features include mean, absolute mean, RMS value, average power, root square amplitude, peak value, peak-to-peak value, variance, standard deviation, skewness, and kurtosis, as well as centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation.

[0087] like Figure 5 As shown, descriptive features are extracted from all data segments, and the core feature set is formed through analysis and processing. The specific steps include the following:

[0088] The time-frequency domain characteristics of meteorological events in the extracted data segments are analyzed to obtain the indicator feature set;

[0089] Based on the dynamic distortion similarity of meteorological events in pairwise data segments, a set of morphological features is obtained;

[0090] The indicator feature set and the morphological feature set are merged to form an initial feature set, which gives the similarity of each meteorological event;

[0091] All meteorological events were clustered using cluster analysis, and pseudo-labels were obtained for each meteorological event.

[0092] Based on the pseudo-labels, a random forest is used to filter the initial feature set to obtain the core feature set of meteorological events.

[0093] Because the values ​​of each dimension have different ranges, the data needs to be transformed into a unitless dataset before use, removing the units.

[0094] The analysis of the time-frequency domain characteristics of meteorological events in the extracted data segments yields an indicator feature set. This also includes dimensionless processing of the indicator feature set data, using the following formula:

[0095] ;

[0096] in, The raw data for the indicator features, for Dimensionless data of indicator characteristics for The mean of the indicator characteristics, for Standard deviation of indicator characteristics.

[0097] The indicator feature set includes a time-domain feature set and a frequency-domain feature set. The time-domain feature set includes waveform indicators, peak values, impulse indicators, margin indicators, skewness indicators, and kurtosis indicators. The frequency-domain feature set includes centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation. The specific relationships between the parameters of the time-domain and frequency-domain feature sets are as follows:

[0098] The relationship formula for the waveform indicators is as follows:

[0099] ;

[0100] The formula for the relationship of peak indicators is as follows:

[0101] ;

[0102] The formula for the relationship of the pulse index is as follows:

[0103] ;

[0104] The formula for the relationship of the margin index is as follows:

[0105] ;

[0106] The formula for the relationship of the skewness index is as follows:

[0107] ;

[0108] The formula for the kurtosis index is as follows:

[0109] ;

[0110] Where n is the number of data segments. This is the data segment.

[0111] In addition to time-frequency domain characteristics, it is also necessary to examine the morphological similarity of each data segment. The indicator feature set and the morphological feature set are merged to form an initial feature set, which gives the similarity of each meteorological event. Specifically, the similarity of each meteorological event is measured by the distance between data segments. The distance relationship between data segments is as follows:

[0112] ;

[0113] in, This represents the distance between two data segments in the time-frequency domain feature space. The cost of morphological transformation between the two data segments. Let n be the number of data segments.

[0114] Based on the core feature set and a portion of the labeled data segments containing high-risk meteorological events, cluster analysis is performed on each high-risk meteorological event, specifically including the following steps:

[0115] Provide the core feature set of a portion of the labeled data segments containing high-risk meteorological events, and obtain the core feature set of high-risk meteorological events;

[0116] The similarity of all core datasets with the core dataset of high-risk meteorological events is compared, and the clustering results of other high-risk meteorological events in the core datasets are given.

[0117] like Figure 6 As shown, cluster analysis was performed on certain data segments, and the clustering results of high-risk meteorological events are presented.

[0118] Acquire the temporal and spatial characteristic parameters of high-risk meteorological events to complete the identification of high-risk meteorological events based on monitoring data from micro-meteorological stations along power transmission lines.

[0119] like Figure 7 As shown, the results are presented in the time dimension. Each meteorological event is classified and statistically analyzed within the category of meteorological events that caused the collapse of power transmission towers. The high-risk meteorological data segments are mainly distributed in April, May, July and August, which is consistent with the frequency of abnormal wind weather, indicating that disaster relief preparations need to be strengthened in the corresponding months.

[0120] The results are presented in spatial dimensions (i.e., latitude and longitude), and a spatial heat map is obtained based on the frequency of high-risk meteorological events corresponding to each transmission line micro-meteorological station. Taking Jiangsu Province as an example, areas with high heat intensity are mainly concentrated in eastern and southern Jiangsu. Eastern Jiangsu is a coastal area with no obstructions at low altitudes over the sea, resulting in less cyclone obstruction and wind speed loss, making it prone to high-risk meteorological events that threaten the power grid system. Southern Jiangsu has lower latitude, and tropical depressions often move from south to north when passing through Jiangsu, making it more likely to be affected by tropical depressions and corresponding to more high-risk meteorological events. The higher the heat intensity areas, the greater the number of high-risk meteorological events, and the higher the safety risks posed to nearby transmission towers. Therefore, it is necessary to strengthen risk assessment and daily maintenance in these areas.

[0121] like Figure 8 As shown, the present invention also provides a high-risk weather identification device based on transmission line micro-weather station monitoring data, employing the high-risk weather identification method based on transmission line micro-weather station monitoring data as described above, including:

[0122] The data acquisition module collects monitoring data from micro-meteorological stations along power transmission lines and segments the data into multiple data segments using time-series data segmentation.

[0123] The annotation module extracts meteorological events from each data segment, and annotates the extracted meteorological events in the data segments according to the high-risk meteorological event marking strategy, thereby obtaining a portion of the data segments containing high-risk meteorological events.

[0124] The analysis and processing module extracts descriptive features from all data segments, analyzes and processes them to form a core feature set, and combines the core feature set with a number of labeled data segments containing high-risk meteorological events to perform cluster analysis on each high-risk meteorological event, obtain the temporal and spatial feature parameters of the high-risk meteorological events, and complete the identification of high-risk meteorological events based on the monitoring data of micro meteorological stations on power transmission lines.

[0125] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for identifying high-risk weather conditions based on monitoring data from micro-meteorological stations along power transmission lines, characterized in that, Includes the following steps: Collect monitoring data from micro-meteorological stations along power transmission lines, and divide the data into multiple data segments by time series segmentation, then extract meteorological events from each data segment; Based on the high-risk meteorological event labeling strategy, meteorological events in the extracted data segments are labeled to obtain a portion of the data segments containing high-risk meteorological events; Descriptive features are extracted from all data segments and analyzed to form a core feature set. These descriptive features include static time-frequency domain features and dynamic distortion similarity features, specifically: The time-frequency domain characteristics of meteorological events in the extracted data segments are analyzed to obtain the indicator feature set; Based on the dynamic distortion similarity of meteorological events in pairwise data segments, a set of morphological features is obtained; The indicator feature set and the morphological feature set are merged to form an initial feature set, which gives the similarity of each meteorological event. Specifically, the similarity of each meteorological event is measured by the distance between data segments. The distance relationship between data segments is as follows: ； in, This represents the distance between two data segments in the time-frequency domain feature space. The cost of morphological transformation between the two data segments. Here, n represents the number of data segments. All meteorological events were clustered using cluster analysis, and pseudo-labels were obtained for each meteorological event. Based on the pseudo-labels, a random forest is used to filter the initial feature set to obtain the core feature set of meteorological events; By combining the core feature set and a portion of the data segments containing high-risk meteorological events that were labeled, cluster analysis was performed on each high-risk meteorological event; Acquire the temporal and spatial characteristic parameters of high-risk meteorological events to complete the identification of high-risk meteorological events based on monitoring data from micro-meteorological stations along power transmission lines.

2. The high-risk meteorological identification method based on transmission line micro-meteorological station monitoring data as described in claim 1, characterized in that, The time series data is segmented into multiple data segments, and meteorological events are extracted from each data segment. The specific steps include the following: The monitoring data from the micro-meteorological stations along the transmission lines for each data segment are set to follow a normal distribution; According to the time series, the monitoring data of micro-meteorological stations on transmission lines are searched for breakpoints, and the final breakpoint set is determined by the maximum threshold of probability likelihood estimation. Based on the breakpoint set, the monitoring data of the micro meteorological station of the transmission line is divided into multiple data segments, and the corresponding meteorological events are identified and extracted in each data segment; The process involves searching for breakpoints in the micro-meteorological station monitoring data of transmission lines according to the time series sequence, and determining the final breakpoint set by using the maximum threshold of probability likelihood estimation. The specific steps are as follows: By using the time series sequence, a combination of breakpoints composed of each initial breakpoint is randomly generated; By performing global and local searches on the monitoring data to update the breakpoints, the final set of breakpoints is determined. The global search generates candidate breakpoints, while the local search performs probability likelihood estimation on each candidate breakpoint to determine the final breakpoint.

3. The high-risk meteorological identification method based on transmission line micro-meteorological station monitoring data as described in claim 1, characterized in that, High-risk weather event labeling strategies include: Obtain the time and latitude / longitude data of the actual disaster event, compare it with the meteorological events in the data segment, and determine the corresponding data segment as the data segment containing high-risk meteorological events.

4. The high-risk meteorological identification method based on transmission line micro-meteorological station monitoring data as described in claim 1, characterized in that, The analysis of the time-frequency domain characteristics of meteorological events in the extracted data segments yields an indicator feature set. This also includes dimensionless processing of the indicator feature set data, using the following formula: ； in, The raw data for the indicator features, for Dimensionless data of indicator characteristics for The mean of the indicator characteristics, for Standard deviation of indicator characteristics.

5. The high-risk meteorological identification method based on transmission line micro-meteorological station monitoring data as described in claim 1, characterized in that, The indicator feature set includes a time domain feature set and a frequency domain feature set. The time domain feature set includes waveform indicators, peak indicators, impulse indicators, margin indicators, skewness indicators, and kurtosis indicators. The frequency domain feature set includes centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation.

6. The high-risk meteorological identification method based on transmission line micro-meteorological station monitoring data as described in claim 1, characterized in that, Based on the core feature set and a portion of the labeled data segments containing high-risk meteorological events, cluster analysis is performed on each high-risk meteorological event, specifically including the following steps: Provide the core feature set of a portion of the labeled data segments containing high-risk meteorological events, and obtain the core feature set of high-risk meteorological events; The similarity of all core datasets with the core dataset of high-risk meteorological events is compared, and the clustering results of other high-risk meteorological events in the core datasets are given.

7. A high-risk weather identification device based on monitoring data from micro-meteorological stations along power transmission lines, characterized in that, The high-risk weather identification method based on micro-meteorological station monitoring data of transmission lines as described in any one of claims 1-6 includes: The data acquisition module collects monitoring data from micro-meteorological stations along power transmission lines and segments the data into multiple data segments using time-series data segmentation. The annotation module extracts meteorological events from each data segment, and annotates the extracted meteorological events in the data segments according to the high-risk meteorological event marking strategy, thereby obtaining a portion of the data segments containing high-risk meteorological events. The analysis and processing module extracts descriptive features from all data segments and analyzes them to form a core feature set. These descriptive features include static time-frequency domain features and dynamic distortion similarity features. Specifically, this includes: analyzing the extracted time-frequency domain features of meteorological events in the data segments to obtain an index feature set; obtaining a morphological feature set based on the dynamic distortion similarity of meteorological events in pairwise data segments; and fusing the index feature set and the morphological feature set to form an initial feature set, providing the similarity of each meteorological event. Specifically, the similarity of each meteorological event is measured by the distance between data segments, and the distance relationships between data segments are as follows: ； in, This represents the distance between two data segments in the time-frequency domain feature space. The cost of morphological transformation between the two data segments. Let n be the number of data segments. All meteorological events are clustered using cluster analysis to obtain pseudo-labels for each event. Random forest is used to filter the initial feature set based on the pseudo-labels, resulting in a core feature set for each meteorological event. Combining the core feature set with a subset of labeled data segments containing high-risk meteorological events, cluster analysis is performed on each high-risk meteorological event to obtain its temporal and spatial characteristic parameters, thus completing the identification of high-risk meteorological events based on transmission line micro-meteorological station monitoring data.

Images