Power transmission tower slope disaster rainfall early warning method and system based on big data analysis

Abstract

The application discloses a power transmission tower slope disaster rainfall early warning method and system based on big data analysis, which comprises the following steps: collecting multi-dimensional data based on the slope area of the power transmission tower, obtaining a multi-dimensional data set and performing structured storage, and constructing a comprehensive database; using big data analysis to mine slope disaster correlation of the comprehensive database, and generating disaster correlation rules; performing regression calculation based on the disaster correlation rules to determine a rainfall influence index; performing rainfall sensitivity analysis on the slope disaster according to the rainfall influence index to obtain an influence sensitivity coefficient; performing rainfall prediction based on the comprehensive database to obtain slope predicted rainfall data; combining the slope predicted rainfall data with the influence sensitivity coefficient to perform rainfall early warning, generating a disaster risk index to activate a multi-level slope disaster rainfall early warning signal to perform real-time early warning on the power transmission tower slope disaster rainfall. The application solves the technical problems of insufficient precision and real-time performance of rainfall early warning.

Description

Technical Field

[0001] This invention relates to the field of rainfall early warning technology, specifically to a rainfall early warning method and system for power transmission tower slope disasters based on big data analysis. Background Technology

[0002] Transmission towers, as an important component of power grid infrastructure, are mostly built in complex terrains, especially traversing mountainous and hilly areas. These areas are prone to landslides, collapses, and other slope disasters under the influence of natural factors such as heavy rainfall, posing a serious threat to the tower structure and even causing major accidents such as tower collapse and line breakage. Rainfall is the main natural factor that induces slope disasters. Especially when extreme weather occurs frequently, the frequency and severity of rainfall-induced slope disasters are increasing. Therefore, how to effectively and accurately predict slope disasters of transmission towers has become crucial.

[0003] Traditional slope disaster early warning relies on rainfall thresholds, neglecting the combined influence of multiple factors such as slope geological conditions, tower structural characteristics, and historical disaster data. This leads to false alarms or missed alarms in early warning signals. Furthermore, it lacks in-depth analysis of the complex relationship between rainfall and slope disasters, making it difficult to adapt to the differences in geological environment and climate characteristics in different regions. This affects the reliability and real-time performance of rainfall early warning for actual power transmission tower slope disasters.

[0004] Therefore, current technologies lack comprehensive analysis of the combined effects of slope geological conditions, tower structural characteristics, and historical disaster data, resulting in insufficient accuracy and real-time performance of rainfall warnings. Summary of the Invention

[0005] The technical problem this invention aims to solve is that traditional slope disaster early warning methods lack comprehensive analysis of the combined effects of slope geological conditions, tower structural characteristics, and historical disaster data, resulting in insufficient accuracy and real-time performance of rainfall warnings. This invention provides a method and system for rainfall early warning of power transmission tower slope disasters based on big data analysis. It collects data from the slope area of ​​the power transmission tower, obtains a multidimensional dataset, stores it in a structured manner, and constructs a comprehensive database. It performs slope disaster correlation mining to determine rainfall impact indicators; it conducts rainfall sensitivity analysis on slope disasters to obtain impact sensitivity coefficients; it obtains predicted rainfall data for the slope and issues rainfall warnings, generating a disaster risk index to activate multi-level slope disaster rainfall warning signals for real-time early warning of rainfall disasters on power transmission tower slopes. This invention solves the technical problem of insufficient accuracy and real-time performance of rainfall warnings due to the lack of comprehensive analysis of the combined effects of slope geological conditions, tower structural characteristics, and historical disaster data in existing technologies. It achieves dynamic hierarchical rainfall warnings, reduces false alarms and missed alarms, and improves the accuracy and real-time performance of rainfall warnings.

[0006] This invention is achieved through the following technical solution:

[0007] In a first aspect, the present invention provides a method for early warning of rainfall-related disasters on power transmission tower slopes based on big data analysis, the method comprising:

[0008] Multidimensional data was collected from the sloping area of ​​the transmission tower to obtain a multidimensional dataset; the multidimensional dataset was then structured and stored to construct a comprehensive database.

[0009] Big data analytics were used to mine slope disaster associations in a comprehensive database and generate disaster association rules. Regression calculations were then performed based on these rules to determine rainfall impact indicators.

[0010] Based on rainfall impact indicators, a rainfall sensitivity analysis of slope disasters is conducted to obtain the impact sensitivity coefficient;

[0011] Rainfall forecasting is performed based on a comprehensive database to obtain slope-predicted rainfall data. The slope-predicted rainfall data is combined with the impact sensitivity coefficient to generate rainfall warnings and generate a disaster risk index to activate multi-level slope disaster rainfall warning signals for real-time warning of rainfall disasters on power transmission tower slopes.

[0012] Furthermore, multidimensional data is collected based on the sloping area of ​​the transmission tower to obtain a multidimensional dataset; and the multidimensional dataset is then structured and stored to construct a comprehensive database, including:

[0013] Key slope points are set in the sloping areas of transmission towers. Key slope points refer to the slope locations that are most prone to deformation or have a decisive impact on the safety of the tower, as determined by geological surveys, historical disaster records, and topographic analysis.

[0014] A monitoring network consisting of multiple types of sensors is deployed at key slope locations to build an Internet of Things (IoT) sensor network.

[0015] Define precipitation meteorological parameters and non-precipitation meteorological parameters, preset a first sampling frequency based on precipitation meteorological parameters, and preset a second sampling frequency based on non-precipitation meteorological parameters;

[0016] The slope of the power transmission tower is synchronously collected in multiple dimensions using an Internet of Things (IoT) sensor network at either a first or a second sampling frequency to obtain a multidimensional dataset. The multidimensional dataset includes multidimensional rainfall data and multidimensional auxiliary data.

[0017] A comprehensive database is constructed by using multidimensional rainfall data and multidimensional auxiliary data as storage indexes according to their acquisition time sequence.

[0018] Furthermore, big data analytics are used to mine slope hazard associations in the comprehensive database, generating hazard association rules, including:

[0019] Samples were extracted from a comprehensive database to obtain an analytical sample set. The analytical sample set includes historical rainfall data sequences and historical slope disaster event records, and there is a correspondence between the historical rainfall data sequences and the historical slope disaster event records.

[0020] Multiple time windows are defined, and cumulative rainfall is calculated by combining historical rainfall data sequences with multiple time windows to obtain cumulative rainfall data for multiple time windows;

[0021] Rainfall intensity is analyzed based on historical rainfall data sequences to obtain the maximum rainfall intensity.

[0022] Feature analysis is performed on the maximum rainfall intensity and the cumulative rainfall data over multiple time windows to generate rainfall characteristic indicators;

[0023] By linking rainfall characteristic indicators with historical slope disaster event records, disaster association rules are generated.

[0024] Furthermore, regression calculations are performed based on disaster association rules to determine rainfall impact indicators, including:

[0025] Rainfall characteristic indicators were used as independent variables, and historical slope disaster event records were used as dependent variables.

[0026] Logistic regression was performed on the analysis sample set based on the independent and dependent variables to obtain the regression coefficients of the rainfall characteristic indicators;

[0027] Rainfall characteristic indicators with absolute values ​​of regression coefficients greater than a preset threshold are selected as the initial indicator set;

[0028] Confidence analysis is performed based on disaster association rules. The disaster association rules are sorted in descending order of confidence to obtain a disaster association rule sequence, from which the first K rules are extracted, where K is an integer greater than 0.

[0029] The high-frequency feature index set is obtained by filtering the rainfall feature indicators based on the first K rules;

[0030] The initial set of indicators and the high-frequency characteristic indicator set are merged and deduplicated to determine the indicators affecting rainfall.

[0031] Furthermore, based on rainfall impact indicators, a rainfall sensitivity analysis was conducted on slope hazards to obtain impact sensitivity coefficients, including:

[0032] Based on the analysis of the sample set, the sample group was divided into a disaster-occurred sample group and a disaster-free sample group;

[0033] The rainfall impact index is mapped to the disaster occurrence sample group for distribution calculation, and the first numerical distribution histogram is constructed.

[0034] The rainfall impact index is mapped to the sample group where no disaster occurred for distribution calculation, and a second numerical distribution histogram is constructed.

[0035] Based on the first and second numerical distribution histograms, the differences in the distribution centers of rainfall impact indicators are calculated to obtain the first difference measure.

[0036] Based on the first and second numerical distribution histograms, the distribution discrete differences of rainfall impact indicators are calculated to obtain a second difference measure.

[0037] Based on the first and second difference measures, multiple quantitative assessments are conducted on the initial sensitivity of slope hazards based on rainfall impact indicators to generate impact sensitivity coefficients.

[0038] Furthermore, based on the comprehensive database, sample extraction is performed to obtain an analytical sample set, including:

[0039] Set the time and geographical scope of the target analysis, traverse the comprehensive database according to the time scope to extract the first slope disaster event of the time scope, and traverse the comprehensive database according to the geographical scope to extract the second slope disaster event of the geographical scope.

[0040] The data of the first slope disaster event and the second slope disaster event are merged to construct a historical record of slope disaster events;

[0041] By traversing historical slope disaster event records and extracting rainfall data from the comprehensive database based on the occurrence time data, a historical rainfall data sequence is obtained.

[0042] Multiple negative samples were obtained by randomly selecting from historical slope disaster event records.

[0043] Based on historical rainfall data sequences, historical slope disaster event records were matched and screened to obtain multiple positive samples;

[0044] Multiple positive samples are combined with multiple negative samples to obtain data combination results; and disasters are identified based on the data combination results to construct an analysis sample set.

[0045] Furthermore, based on the correlation mining between rainfall characteristic indicators and historical slope disaster event records, disaster correlation rules are generated, including:

[0046] The rainfall characteristic indicators are numerically discretized to determine several level intervals. Based on these level intervals, the rainfall characteristic indicators are transformed to generate independent terms. Historical slope disaster event records are treated as special terms.

[0047] Combine independent items with special items to construct a transaction item set;

[0048] The association rule mining algorithm is used to scan the transaction itemset, count the first occurrence frequency of each item, and select items whose first occurrence frequency is not lower than the preset minimum support threshold to form frequent 1-itemsets;

[0049] Join frequent 1-itemsets to obtain joined itemsets; prune the joined itemsets to generate candidate 2-itemsets;

[0050] Scan the transaction itemset to count the second occurrence frequency of candidate 2-itemsets, filter candidate 2-itemsets whose second occurrence frequency is not lower than the preset minimum support threshold, and filter out frequent 2-itemsets;

[0051] Iterate through the above process, generating candidate h-itemsets layer by layer and filtering out frequent h-itemsets until no new frequent itemsets can be generated, and construct disaster association rules.

[0052] Furthermore, based on a comprehensive database, rainfall forecasting is performed to obtain slope-predicted rainfall data; this slope-predicted rainfall data is combined with an impact sensitivity coefficient to generate rainfall warnings, creating a disaster risk index to activate multi-level slope disaster rainfall warning signals for real-time early warning of transmission tower slope disasters, including:

[0053] Based on historical and real-time meteorological data accumulated in a comprehensive database, weather forecasting models are invoked to predict rainfall and obtain slope-predicted rainfall data.

[0054] The predicted rainfall data for slopes is combined with the impact sensitivity coefficient for weighted fusion calculation to generate a disaster risk index.

[0055] Multiple consecutive warning threshold intervals are set, each interval corresponding to a risk level; the disaster risk index is compared and judged with multiple consecutive warning threshold intervals, and a target warning level is generated by matching;

[0056] Activate multi-level slope disaster rainfall warning signals based on the target warning level;

[0057] Multi-level slope disaster rainfall warning signals are synchronized to remote operation and maintenance terminals for disaster rainfall alarms;

[0058] Among them, the multi-level slope disaster rainfall warning signal includes a standardized response plan corresponding to each level. Different levels of slope disaster rainfall warning signals have different warning names, different response measures, and different information alarm ranges.

[0059] Furthermore, the predicted rainfall data for slopes is combined with the impact sensitivity coefficient for weighted fusion calculation to generate a disaster risk index, including:

[0060] By combining slope-predicted rainfall data with impact sensitivity coefficients, multiple rainfall impact sensitivity indicators are obtained; these multiple rainfall impact sensitivity indicators include cumulative rainfall sensitivity indicators, average rainfall intensity sensitivity indicators, and maximum rainfall intensity sensitivity indicators.

[0061] Risk calculations are performed based on cumulative rainfall, average rainfall intensity, and maximum rainfall intensity sensitive indicators to obtain multiple sub-risk values.

[0062] The cumulative rainfall sensitive index, the average rainfall intensity sensitive index, and the maximum rainfall intensity sensitive index are weighted and integrated according to multiple sub-risk values ​​to obtain multiple weighted risk contribution degrees.

[0063] A disaster risk index is generated by comprehensively calculating multiple rainfall-sensitive indicators and multiple weighted risk contribution factors.

[0064] Secondly, this invention also provides a rainfall-based early warning system for power transmission tower slope disasters based on big data analysis. This system includes:

[0065] The data acquisition unit is used to collect multidimensional data based on the sloping area of ​​the transmission tower to obtain a multidimensional dataset; and to store the multidimensional dataset in a structured manner to build a comprehensive database.

[0066] The rainfall impact index determination unit is used to perform slope disaster correlation mining on the comprehensive database using big data analysis to generate disaster correlation rules; and to perform regression calculations based on the disaster correlation rules to determine rainfall impact indicators.

[0067] The rainfall sensitivity analysis unit is used to perform rainfall sensitivity analysis on slope disasters based on rainfall impact indicators and obtain the impact sensitivity coefficient.

[0068] The rainfall early warning unit is used to predict rainfall based on a comprehensive database and obtain slope predicted rainfall data; it combines the slope predicted rainfall data with the influence sensitivity coefficient to generate a disaster risk index to activate multi-level slope disaster rainfall early warning signals and provide real-time early warning of slope disasters to power transmission towers.

[0069] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0070] This invention presents a method and system for early warning of rainfall-related disasters on power transmission tower slopes based on big data analysis. It collects data from the slope areas of power transmission towers, obtains multidimensional datasets, stores them in a structured manner, and constructs a comprehensive database. The comprehensive database is then used for slope disaster correlation mining to determine rainfall impact indicators. Rainfall sensitivity analysis is performed on slope disasters to obtain impact sensitivity coefficients. Predicted rainfall data for slopes is obtained and rainfall warnings are issued. A disaster risk index is generated to activate multi-level slope disaster rainfall warning signals, providing real-time early warnings for rainfall-related disasters on power transmission tower slopes. This invention solves the technical problem in existing technologies where the lack of analysis of the comprehensive impact of slope geological conditions, tower structural characteristics, and historical disaster data leads to insufficient accuracy and real-time performance of rainfall warnings. It achieves dynamic, tiered rainfall warnings, reduces false alarms and missed alarms, and improves the accuracy and real-time performance of rainfall warnings. Attached Figure Description

[0071] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0072] Figure 1 This is a flowchart of the rainfall early warning method for power transmission tower slope disasters based on big data analysis, as described in this invention.

[0073] Figure 2 This is a structural block diagram of the power transmission tower slope disaster rainfall early warning system based on big data analysis according to the present invention. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0075] Example 1

[0076] like Figure 1 As shown, this invention provides a rainfall-based early warning method for power transmission tower slope disasters based on big data analysis. The method includes:

[0077] Step S100: Multidimensional data is collected based on the slope area of ​​the transmission tower to obtain a multidimensional dataset; and the multidimensional dataset is stored in a structured manner to construct a comprehensive database;

[0078] In this embodiment, step S100 specifically includes:

[0079] S110, set key slope points in the sloping area of ​​the transmission tower;

[0080] S120 deploys a monitoring network consisting of various types of sensors at key slope locations to build an Internet of Things sensor network;

[0081] S130, define precipitation meteorological parameters and non-precipitation meteorological parameters, preset a first sampling frequency based on precipitation meteorological parameters, and preset a second sampling frequency based on non-precipitation meteorological parameters;

[0082] S140, the slope of the power transmission tower is synchronously collected in multiple dimensions through the Internet of Things sensor network at either the first sampling frequency or the second sampling frequency to obtain a multidimensional dataset; the multidimensional dataset includes multidimensional rainfall data and multidimensional auxiliary data.

[0083] S150 uses multidimensional rainfall data and multidimensional auxiliary data as storage indexes according to the collection time sequence to build a comprehensive database.

[0084] Preferably, the key slope locations refer to the slope positions most prone to deformation or having a decisive impact on the safety of the transmission tower, as determined by geological surveys, historical disaster records, and topographic analysis; such as the slope toe, slope top, geologically vulnerable zones, and areas with existing micro-cracks. Then, a monitoring network composed of various types of sensors is deployed at the key slope locations of the transmission tower, including multiple sensors with different functions that comprehensively perceive the environmental conditions. These may include sensors such as tipping bucket rain gauges and laser rain spectrometers to collect rainfall data, sensors such as soil moisture sensors, inclinometers, crack gauges, and strain gauges to collect slope condition data, and sensors such as anemometers, temperature and humidity sensors, and barometers to collect environmental data. Furthermore, multiple sensors are interconnected through wireless communication modules and network protocols.

[0085] Preferably, precipitation meteorological parameters and non-precipitation meteorological parameters are defined. Precipitation meteorological parameters refer to data directly related to precipitation, mainly rainfall amount and rainfall intensity. Non-precipitation meteorological parameters refer to other data that change relatively slowly or are not timely related to disaster triggering, such as temperature, humidity, wind speed, air pressure, and geological parameters such as soil volumetric water content and slope inclination angle. A higher frequency is preset for precipitation meteorological parameters as the first sampling frequency, such as once per minute or once every 5 minutes, to accurately capture rainfall intensity. A relatively lower frequency is preset for non-precipitation meteorological parameters as the second sampling frequency, such as once per hour or once every 30 minutes, to significantly save sensor power consumption and reduce network data transmission volume.

[0086] Preferably, the slope of the transmission tower is synchronously collected in multiple dimensions using an IoT sensor network at either a first or second sampling frequency. This ensures that all different types of data are recorded under a unified timestamp, including multiple dimensions such as multidimensional rainfall data and multidimensional auxiliary data. These are combined to form a multidimensional dataset. Each data point includes a timestamp and corresponding collected data such as rainfall, soil moisture content, tilt angle, and wind speed. Soil moisture content, tilt angle, and wind speed are considered auxiliary data. Then, the multidimensional rainfall data and multidimensional auxiliary data are stored as storage indexes according to the collection time sequence to facilitate efficient querying. This leads to the determination of a comprehensive database for storing and managing all raw time-series data from the IoT sensor network.

[0087] Step S200: Use big data analysis to mine slope disaster associations in the comprehensive database and generate disaster association rules; perform regression calculations based on the disaster association rules to determine rainfall impact indicators;

[0088] In this embodiment, step S200 includes:

[0089] S210, Sample extraction is performed based on a comprehensive database to obtain an analysis sample set; the analysis sample set includes historical rainfall data sequences and historical slope disaster event records, and there is a correspondence between the historical rainfall data sequences and the historical slope disaster event records;

[0090] S220: Multiple time windows are defined, and cumulative rainfall is calculated based on historical rainfall data sequences combined with multiple time windows to obtain cumulative rainfall data for multiple time windows;

[0091] S230, based on historical rainfall data series, performs rainfall intensity analysis to obtain the maximum rainfall intensity;

[0092] S240 performs feature analysis on the maximum rainfall intensity and the cumulative rainfall data of multiple time windows to generate rainfall feature indicators;

[0093] S250 generates disaster association rules by linking rainfall characteristic indicators with historical slope disaster event records.

[0094] Preferably, big data analytics is used to mine slope disaster associations in a comprehensive database to generate disaster association rules. Specifically, samples are extracted from the comprehensive database to form an analysis sample set, which includes historical rainfall data sequences and historical slope disaster event records. The historical rainfall data sequence refers to the original rainfall data extracted from each sample after a period of time back to its corresponding time, such as the rainfall per minute / hour. The historical slope disaster event records are used to indicate the slope disaster that occurred at each specific time and location, and are usually binary labels, such as 1 indicating that it occurred and 0 indicating that it did not occur. Moreover, there is a correspondence between the historical rainfall data sequence and the historical slope disaster event records, that is, each historical rainfall data sequence is spatiotemporally aligned with the historical slope disaster event records. For example, for a positive disaster sample, the rainfall data sequence for a period of time before the disaster occurs is extracted. For a negative disaster sample, a date when no disaster occurred is randomly selected, and a rainfall data sequence of the same length is extracted.

[0095] Preferably, different types of slope hazards may be induced by rainfall of different durations. For example, shallow landslides are sensitive to short-duration heavy rainfall, while deep landslides are sensitive to long-term cumulative rainfall. Therefore, multiple time windows are defined, i.e., different backtracking periods are pre-set, such as 30 minutes, 1 hour, 3 hours, 12 hours, and 48 hours. For each historical rainfall data sequence, the total rainfall of multiple time windows is backtracked from the end time of the sequence to obtain the cumulative rainfall data for multiple time windows. Simultaneously, rainfall intensity analysis is performed based on the historical rainfall data sequence, i.e., the maximum rainfall per unit time within the time range of the historical rainfall data sequence is calculated, i.e., the maximum rainfall intensity is obtained, which is used to capture short-duration rainstorm peaks that may lead to slope hazards. Next, feature analysis is performed on the maximum rainfall intensity and the cumulative rainfall data from multiple time windows, combining them into a rainfall feature vector to describe the overall picture of a rainfall process, thereby generating rainfall feature indicators. Finally, an association mining algorithm is applied to perform association mining based on the rainfall feature indicators and historical slope disaster event records. First, the continuous rainfall feature indicators are discretized, and then the historical slope disaster event records are scanned to identify frequently occurring discretized rainfall feature indicators and labels indicating whether a disaster has occurred, thereby generating disaster association rules. Among them, disaster association rules have support and confidence. Confidence represents the probability of the combination within the rule occurring. The higher the confidence, the more reliable the rule. Support represents the frequency of the rule's occurrence in all samples.

[0096] Further, step S210 includes:

[0097] S211, Set the time range and geographical range of the target analysis, traverse the comprehensive database according to the time range to extract the first slope disaster event of the time range, and traverse the comprehensive database according to the geographical range to extract the second slope disaster event of the geographical range.

[0098] S212, merge the data of the first slope disaster event and the second slope disaster event to construct a historical slope disaster event record;

[0099] S213, Traverse the historical slope disaster event records and extract rainfall data from the comprehensive database according to the occurrence time data to obtain the historical rainfall data sequence;

[0100] S214, multiple negative samples were obtained by randomly selecting from historical slope disaster event records;

[0101] S215, based on historical rainfall data sequences, matched and screened historical slope disaster event records to obtain multiple positive samples;

[0102] S216 combines multiple positive samples with multiple negative samples to obtain data combination results; and identifies disasters based on the data combination results to construct an analysis sample set.

[0103] Preferably, the time and geographical scope of the target analysis are set to clarify the boundaries of data screening. For example, data on the slopes of all transmission towers located in mountainous and hilly areas over the past five years are screened. Then, the comprehensive database is traversed according to the time and geographical scope respectively to extract the first slope disaster events of the time scope and the second slope disaster events of the geographical scope. These include all disaster events that occurred within the target time scope and all disaster events that occurred within the target geographical scope, respectively. The data of the first slope disaster events and the second slope disaster events are merged to construct a historical slope disaster event record. For each record in the historical slope disaster event record, based on its occurrence time data and corresponding occurrence location, rainfall data for a period of time before the disaster is extracted from the comprehensive database, thereby generating a historical rainfall data sequence to describe the specific rainfall process of this slope disaster.

[0104] Preferably, sample data of disasters are randomly extracted from historical slope disaster event records. That is, within the target geographical area, multiple tower slope locations where disasters have occurred are randomly selected, and within the target time range, time points of safe locations are randomly selected to obtain multiple negative samples. Each negative sample represents a disaster event that occurred at a specific time and a specific safe location. Then, the historical slope disaster event records are matched and filtered according to the historical rainfall data sequence. That is, the event records of slope disasters that occurred within the historical rainfall data sequence are used as positive samples. Each positive sample represents a confirmed disaster that occurred at a specific time and location and has complete pre-disaster rainfall data. Finally, multiple positive samples and multiple negative samples are combined to obtain the data combination result and disaster labeling is performed. That is, a disaster label is added to each sample. The label for positive samples is 1, representing slope disasters caused by rainfall, and the label for negative samples is 0, representing disasters that occurred without rainfall. Finally, an analysis sample set is constructed.

[0105] Further, step S250 includes:

[0106] S251, the rainfall characteristic indicators are numerically discretized to determine several level intervals, and the rainfall characteristic indicators are transformed based on several level intervals to generate independent terms; and historical slope disaster event records are treated as special terms;

[0107] S252, combine independent items and special items to construct a transaction item set;

[0108] S253, use association rule mining algorithm to scan transaction itemsets, count the first occurrence frequency of each item, and filter items whose first occurrence frequency is not lower than the preset minimum support threshold to form frequent 1-itemsets;

[0109] S254: Perform joins based on frequent 1-itemsets to obtain joined itemsets; prune based on joined itemsets to generate candidate 2-itemsets;

[0110] S255, scan the transaction itemset to count the second occurrence frequency of candidate 2-itemsets, filter candidate 2-itemsets whose second occurrence frequency is not lower than the preset minimum support threshold, and filter out frequent 2-itemsets;

[0111] S256, iterate the above process, generate candidate h-itemsets layer by layer and filter out frequent h-itemsets until no new frequent itemsets can be generated, and construct disaster association rules.

[0112] Preferably, the rainfall characteristic indicators are numerically discretized, converting continuous rainfall characteristic indicators into categorical variables. Several level intervals are determined for each rainfall characteristic, and then the rainfall characteristic indicators are transformed according to these level intervals to generate independent items. Historical slope disaster event records are then used as special items. Independent items and special items are combined to construct transaction itemsets. For samples where a disaster occurred, the transaction itemset includes the discretized results of all rainfall characteristic indicators corresponding to that sample and the special itemsets. For samples where no disaster occurred, the transaction itemset only includes the discretized results of all rainfall characteristic indicators corresponding to that sample. A minimum support threshold is set based on historical data and experience to measure the frequency of a pattern. Support is the ratio of the number of transactions containing a certain itemset to the total number of transactions. Then, an association rule mining algorithm is used to scan the transaction itemsets, counting the frequency of each item to determine the first occurrence frequency. All items with a first occurrence frequency not lower than the minimum support threshold are selected to form frequent 1-itemsets.

[0113] Preferably, the method involves joining based on frequent 1-itemsets, pruning based on the joined itemsets, that is, combining all items in the frequent 1-itemsets in pairs to generate all possible 2-itemets, generating candidate 2-itemsets, then checking whether the subsets of the candidate 2-itemsets are all in the frequent 1-itemsets. If a subset is infrequent, it is pruned. Then, the transaction itemsets are scanned to count the second occurrence frequency of the candidate 2-itemsets, and candidate 2-itemsets with a second occurrence frequency not lower than the minimum support threshold are selected to identify frequent 2-itemsets. The method is recursively used to discover larger-scale frequent patterns, generating candidate h-itemsets layer by layer and selecting frequent h-itemsets until no new frequent itemsets can be generated. For each frequent itemset, multiple rules are generated and the confidence is calculated. Finally, rules are generated, including all strongly associated rules whose support and confidence reach the preset threshold.

[0114] Furthermore, step S200 also includes:

[0115] S260 uses rainfall characteristic indicators as independent variables and historical slope disaster event records as dependent variables;

[0116] S270, Logistic regression is performed on the analysis sample set based on independent and dependent variables to obtain the regression coefficients of rainfall characteristic indicators;

[0117] S280: Select rainfall characteristic indicators whose absolute values ​​of regression coefficients are greater than a preset threshold as the initial set of indicators;

[0118] S290, perform confidence analysis based on disaster association rules, sort the disaster association rules in descending order of confidence, obtain the disaster association rule sequence, and extract the first K rules from it, where K is an integer greater than 0;

[0119] S2100, based on the first K rules, filters the rainfall characteristic indicators to obtain a set of high-frequency characteristic indicators;

[0120] S2110, merge and deduplicate the initial set of indicators and the set of high-frequency characteristic indicators to determine the indicators affecting rainfall.

[0121] Preferably, regression calculations are performed based on disaster association rules. Specifically, rainfall characteristic indicators are used as independent variables, and historical slope disaster event records are used as dependent variables, i.e., a binary variable indicating whether a disaster has occurred. Logistic regression is performed on the analysis sample set according to the independent and dependent variables to predict the probability of disaster events and determine the degree and direction of influence of each independent variable on the dependent variable. That is, the logistic regression model calculates regression coefficients for each independent variable. The absolute value of the regression coefficient directly reflects the strength of the influence of the rainfall characteristic indicator on the probability of disaster occurrence. The larger the absolute value of the regression coefficient, the greater the influence of the rainfall characteristic indicator on the probability of disaster occurrence. A preset threshold is set according to business knowledge and statistical significance level to filter rainfall characteristic indicators with weak influence, and select those with a regression coefficient absolute value greater than the preset threshold. Rainfall characteristic indicators are selected to identify candidate indicators that significantly contribute to disaster prediction as an initial indicator set. Next, a confidence analysis is performed on disaster association rules, and these rules are sorted in descending order of confidence. The top K rules with the highest confidence are selected, representing the strongest and most reliable rainfall pattern-disaster relationship, where K is a positive integer (e.g., K=10). Then, rainfall characteristic indicators are further filtered based on these top K rules to obtain the rainfall characteristic indicators that frequently appear among the many high-confidence rules, resulting in a high-frequency characteristic indicator set. This set represents the key components constituting high-risk rainfall patterns. Finally, the initial indicator set and the high-frequency characteristic indicator set are merged and deduplicated to obtain the final set of rainfall characteristic indicators, which serves as rainfall impact indicators, thereby ensuring the accuracy and reliability of rainfall disaster early warning.

[0122] Step S300: Perform rainfall sensitivity analysis on slope disasters based on rainfall impact indicators to obtain impact sensitivity coefficients;

[0123] In this embodiment, step S300 specifically includes:

[0124] S310, based on the analysis sample set, divide the disaster-occurring sample group and the disaster-free sample group;

[0125] S320 maps the rainfall impact index to the disaster occurrence sample group for distribution calculation and constructs the first numerical distribution histogram;

[0126] S330 maps the rainfall impact index to the sample group where no disaster occurred for distribution calculation and constructs a second numerical distribution histogram.

[0127] S340, based on the first and second numerical distribution histograms, calculate the difference in distribution centers of rainfall impact indicators to obtain the first difference measure;

[0128] S350, based on the first and second numerical distribution histograms, calculates the discrete differences in the distribution of rainfall impact indicators to obtain a second difference measure;

[0129] S360, based on the first and second difference measures, performs multiple quantitative assessments of the initial sensitivity of slope hazards based on rainfall impact indicators, generating an impact sensitivity coefficient.

[0130] Preferably, a rainfall sensitivity analysis of slope disasters is conducted based on rainfall impact indicators. Specifically, the analysis sample set is divided into a disaster-occurring sample group and a disaster-free sample group. The rainfall impact indicators are then mapped to the disaster-occurring sample group and the disaster-free sample group for distribution calculation. That is, the values ​​of each rainfall impact indicator in all samples in the disaster-occurring sample group and the disaster-free sample group are calculated. A first value distribution histogram and a second value distribution histogram are plotted to divide the value range into continuous intervals and count the number of samples in each interval, so as to intuitively display the distribution pattern, central location and dispersion of each rainfall impact indicator.

[0131] Preferably, based on the first and second numerical distribution histograms, the distribution center difference of the rainfall impact index is calculated, i.e., the standardized mean difference of the rainfall impact index is calculated to determine the first difference measure, which is used to quantify the degree of deviation between the disaster group and the non-disaster group in the central position of the rainfall impact index. If the first difference measure value of the rainfall impact index is large, it indicates that the rainfall impact index is significantly higher than the typical value when the disaster occurs. Based on the first and second numerical distribution histograms, the distribution dispersion difference of the rainfall impact index is calculated, i.e., the variance ratio of the rainfall impact index is calculated to determine the second difference measure, which is used to quantify the difference in data volatility or dispersion between the disaster group and the non-disaster group in the rainfall impact index. Finally, based on the first difference measure, the second difference measure, and the rainfall impact index, the initial sensitivity of the slope disaster is quantitatively assessed in multiple ways. That is, the reciprocals of the first and second difference measures of each rainfall impact index are multiplied to obtain the impact sensitivity coefficient of the rainfall impact index, which comprehensively reflects the comprehensive ability of the rainfall impact index to distinguish between disaster and non-disaster states. The larger the impact sensitivity coefficient, the higher the sensitivity of the rainfall impact index to the occurrence of disaster, thereby ensuring the accuracy of the early warning.

[0132] Step S400: Based on the comprehensive database, rainfall prediction is performed to obtain slope predicted rainfall data; the slope predicted rainfall data is combined with the influence sensitivity coefficient to perform rainfall early warning, and a disaster risk index is generated to activate multi-level slope disaster rainfall early warning signals to provide real-time early warning of rainfall disasters on power transmission tower slopes.

[0133] Preferably, based on historical and real-time meteorological data accumulated in a comprehensive database, a weather forecasting model is invoked to predict rainfall. Specifically, for each transmission tower slope location, rainfall data corresponding to rainfall impact indicators is predicted for a future period, outputting the slope's predicted rainfall data. Then, a rainfall warning is generated by combining the slope's predicted rainfall data as input and the impact sensitivity coefficient as weight. This generates a disaster risk index, used to quantify the contribution of different rainfall characteristics to slope disasters, thus more accurately reflecting future disaster risks. Finally, a multi-level slope disaster rainfall warning signal is activated to provide real-time warnings for transmission tower slope disasters. This involves comparing the disaster risk index with multiple preset continuous warning threshold intervals to automatically determine the current target warning level (e.g., blue, yellow, orange, and red). Based on the determined target warning level, the multi-level slope disaster rainfall warning signal is automatically triggered. Each level of the slope disaster rainfall warning signal corresponds to a preset emergency response procedure. The warning signal is then pushed in real-time to the monitoring terminal of remote maintenance personnel via a communication network, achieving disaster rainfall alerts and ensuring improved accuracy and real-time performance of rainfall warnings.

[0134] In this embodiment, step S400 includes:

[0135] S410, based on historical and real-time meteorological data accumulated in the comprehensive database, calls the weather forecasting model to predict rainfall and obtain slope predicted rainfall data;

[0136] S420 combines slope-predicted rainfall data with impact sensitivity coefficients for weighted fusion calculation to generate a disaster risk index.

[0137] S430 sets multiple continuous early warning threshold intervals, each interval corresponding to a risk level; the disaster risk index is compared and judged with multiple continuous early warning threshold intervals to generate a target early warning level;

[0138] S440 activates multi-level slope disaster rainfall warning signals based on the target warning level;

[0139] The S450 synchronizes multi-level slope disaster rainfall warning signals to the remote operation and maintenance terminal for disaster rainfall alarm;

[0140] Among them, the multi-level slope disaster rainfall warning signal includes a standardized response plan corresponding to each level. Different levels of slope disaster rainfall warning signals have different warning names, different response measures, and different information alarm ranges.

[0141] Preferably, the impact sensitivity coefficient is used as a weighting factor to weight and fuse the slope prediction rainfall data and impact sensitivity coefficient for each rainfall impact indicator to generate a disaster risk index, which is a continuous value between 0 and 1. This index is used to quantify the comprehensive risk probability of a disaster occurring on the slope of the transmission tower in the future. Multiple continuous warning threshold intervals are predefined, each interval corresponding to a risk level, such as blue (low risk), yellow (medium risk), orange (high risk), and red (extremely high risk). The disaster risk index is then compared with the multiple continuous warning threshold intervals to determine and generate a target warning level. Multi-level slope disaster rainfall warning signals are then activated based on the target warning level. These signals include standardized response plans for each level, with different warning names, response measures, and information alert ranges for different levels. Finally, the multi-level slope disaster rainfall warning signals are synchronized to remote operation and maintenance terminals for disaster rainfall alerts, i.e., sent to the smart terminals of remote operation and maintenance personnel. This effectively improves the response speed and resource allocation efficiency for disaster prevention and mitigation, and ensures the accuracy and real-time nature of rainfall warnings.

[0142] Furthermore, the S420 specifically includes:

[0143] S421 combines the predicted rainfall data of the slope with the impact sensitivity coefficient for matching calculation to obtain multiple rainfall impact sensitivity indicators; these multiple rainfall impact sensitivity indicators include cumulative rainfall sensitivity indicators, average rainfall intensity sensitivity indicators, and maximum rainfall intensity sensitivity indicators.

[0144] S422 calculates risk based on cumulative rainfall sensitivity index, average rainfall intensity sensitivity index and maximum rainfall intensity sensitivity index to obtain multiple sub-risk values;

[0145] S423 calculates multiple weighted risk contributions by weighting and integrating the cumulative rainfall sensitive index, the average rainfall intensity sensitive index, and the maximum rainfall intensity sensitive index according to multiple sub-risk values.

[0146] S424 combines multiple rainfall-sensitive indicators with multiple weighted risk contribution factors to generate a disaster risk index.

[0147] Preferably, the predicted rainfall data for the slope is matched with the impact sensitivity coefficient for calculation. This involves calculating the product of the predicted rainfall data and the impact sensitivity coefficient for each rainfall impact indicator, identifying multiple rainfall impact sensitivity indicators. Each rainfall impact sensitivity indicator is a risk factor weighted and adjusted from the cumulative rainfall sensitivity indicator, the average rainfall intensity sensitivity indicator, and the maximum rainfall intensity sensitivity indicator. This transforms the rainfall prediction data into values ​​reflecting the disaster risk. Then, risk calculation is performed based on the cumulative rainfall sensitivity indicator, the average rainfall intensity sensitivity indicator, and the maximum rainfall intensity sensitivity indicator. This involves mapping each rainfall impact sensitivity indicator to a linear normalization function to obtain multiple sub-risk values, representing a single rainfall... The risk score contributed by rainfall characteristics is calculated by converting the value of the cumulative rainfall sensitivity index into a cumulative rainfall sub-item risk value between 0 and 1. Then, the cumulative rainfall sensitivity index, average rainfall intensity sensitivity index, and maximum rainfall intensity sensitivity index are weighted and fused according to multiple sub-item risk values. The weights are set based on historical data experience and forecasting needs to obtain multiple weighted risk contribution degrees, representing the share of each sub-item risk in the final risk index. Finally, the multiple rainfall impact sensitivity indices and multiple weighted risk contribution degrees are summed to obtain the disaster risk index, which comprehensively reflects the disaster risk level caused by the combined effect of all key rainfall characteristic indicators and their sensitivities.

[0148] Example 2

[0149] like Figure 2 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment further provides a rainfall-based early warning system for transmission tower slope disasters based on big data analysis. This system corresponds one-to-one with the rainfall-based early warning method for transmission tower slope disasters based on big data analysis in Embodiment 1. The system includes:

[0150] The data acquisition unit is used to collect multidimensional data based on the sloping area of ​​the transmission tower to obtain a multidimensional dataset; and to store the multidimensional dataset in a structured manner to build a comprehensive database.

[0151] The rainfall impact index determination unit is used to perform slope disaster correlation mining on the comprehensive database using big data analysis to generate disaster correlation rules; and to perform regression calculations based on the disaster correlation rules to determine rainfall impact indicators.

[0152] The rainfall sensitivity analysis unit is used to perform rainfall sensitivity analysis on slope disasters based on rainfall impact indicators and obtain the impact sensitivity coefficient.

[0153] The rainfall early warning unit is used to predict rainfall based on a comprehensive database and obtain slope predicted rainfall data; it combines the slope predicted rainfall data with the influence sensitivity coefficient to generate a disaster risk index to activate multi-level slope disaster rainfall early warning signals and provide real-time early warning of slope disasters to power transmission towers.

[0154] The execution process of each unit can be carried out according to the steps of the method for early warning of rainstorm disasters on power transmission tower slopes based on big data analysis in Example 1, and will not be described in detail in this example.

[0155] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0156] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0157] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0158] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0159] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A rainfall-based early warning method for power transmission tower slope disasters based on big data analysis, characterized in that, The method includes: Multidimensional data is collected from the sloping area of ​​the transmission tower to obtain a multidimensional dataset; the multidimensional dataset is then structured and stored to construct a comprehensive database. Big data analytics are used to mine slope disaster associations in the comprehensive database and generate disaster association rules; regression calculations are then performed based on these disaster association rules to determine rainfall impact indicators. Based on the rainfall impact indicators, a rainfall sensitivity analysis was conducted on slope disasters to obtain the impact sensitivity coefficient; Rainfall prediction is performed based on the comprehensive database to obtain slope predicted rainfall data; the slope predicted rainfall data is combined with the influence sensitivity coefficient to generate rainfall warning, and a disaster risk index is generated to activate multi-level slope disaster rainfall warning signals to provide real-time warning of rainfall disasters on power transmission tower slopes.

2. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 1, characterized in that, Multidimensional data is collected from the sloping area of ​​the transmission tower to obtain a multidimensional dataset; the multidimensional dataset is then structured and stored to construct a comprehensive database, including: Key slope points are set in the sloping area of ​​the transmission tower. The key slope points refer to the slope locations that are most prone to deformation or have an impact on the safety of the tower, as determined by geological surveys, historical disaster records, and topographic analysis. A monitoring network consisting of various types of sensors was deployed at the key slope locations to build an Internet of Things (IoT) sensor network. Define precipitation meteorological parameters and non-precipitation meteorological parameters, preset a first sampling frequency based on the precipitation meteorological parameters, and preset a second sampling frequency based on the non-precipitation meteorological parameters; The IoT sensor network is used to synchronously collect multidimensional data on the slope of the power transmission tower according to the first sampling frequency or the second sampling frequency to obtain a multidimensional dataset; the multidimensional dataset includes multidimensional rainfall data and multidimensional auxiliary data. The multidimensional rainfall data and the multidimensional auxiliary data are used as storage indexes according to the collection time sequence to construct a comprehensive database.

3. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 1, characterized in that, Big data analytics are used to mine slope hazard associations in the comprehensive database, generating hazard association rules, including: Based on the comprehensive database, a sample set for analysis is obtained; the sample set for analysis includes historical rainfall data sequences and historical slope disaster event records, and there is a corresponding relationship between the historical rainfall data sequences and the historical slope disaster event records. Multiple time windows are defined, and cumulative rainfall is calculated based on the historical rainfall data sequence and the multiple time windows to obtain the cumulative rainfall data for the multiple time windows; Based on the historical rainfall data sequence, rainfall intensity analysis was performed to obtain the maximum rainfall intensity. The maximum rainfall intensity and the cumulative rainfall data of the multiple time windows are subjected to feature analysis to generate rainfall feature indicators; Based on the rainfall characteristic indicators and the historical slope disaster event records, correlation mining is performed to generate disaster association rules.

4. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 3, characterized in that, Regression calculations are performed based on the aforementioned disaster association rules to determine rainfall impact indicators, including: The rainfall characteristic index is used as the independent variable, and the historical slope disaster event record is used as the dependent variable; Logistic regression is performed on the analysis sample set based on the independent and dependent variables to obtain the regression coefficients of the rainfall characteristic indicators; Rainfall characteristic indicators whose absolute values ​​of regression coefficients are greater than a preset threshold are selected as the initial indicator set; Based on the disaster association rules, a confidence analysis is performed. The disaster association rules are then sorted in descending order of confidence to obtain a disaster association rule sequence from which the first K rules are extracted, where K is an integer greater than 0. The rainfall characteristic indicators are filtered based on the first K rules to obtain a set of high-frequency characteristic indicators; The initial set of indicators and the high-frequency feature indicator set are merged and deduplicated to determine the indicators affecting rainfall.

5. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 3, characterized in that, Based on the aforementioned rainfall impact indicators, a rainfall sensitivity analysis of slope hazards is conducted to obtain impact sensitivity coefficients, including: Based on the analysis sample set, the sample group was divided into a disaster-occurred sample group and a disaster-free sample group; The rainfall impact index is mapped to the disaster occurrence sample group for distribution calculation, and a first numerical distribution histogram is constructed. The rainfall impact index is mapped to the sample group where the disaster did not occur for distribution calculation, and a second numerical distribution histogram is constructed. Based on the first numerical distribution histogram and the second numerical distribution histogram, the difference in distribution center of the rainfall impact index is calculated to obtain a first difference measure; Based on the first numerical distribution histogram and the second numerical distribution histogram, the distribution discrete difference of the rainfall impact index is calculated to obtain a second difference measure; Based on the first difference measure and the second difference measure, the initial sensitivity of slope disasters is quantitatively assessed using the rainfall impact index, generating an impact sensitivity coefficient.

6. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 3, characterized in that, Based on the comprehensive database, sample extraction is performed to obtain an analytical sample set, including: Set the time range and geographical range for target analysis, traverse the comprehensive database according to the time range to extract the first slope disaster event within the time range, and traverse the comprehensive database according to the geographical range to extract the second slope disaster event within the geographical range; The data of the first slope disaster event and the second slope disaster event are merged to construct a historical slope disaster event record; By traversing the historical slope disaster event records and extracting rainfall data from the comprehensive database according to the occurrence time data, a historical rainfall data sequence is obtained. Multiple negative samples were obtained by randomly selecting from the historical slope disaster event records. Based on the historical rainfall data sequence, historical slope disaster event records were matched and screened to obtain multiple positive samples; Multiple positive samples are combined with multiple negative samples to obtain data combination results; and disasters are identified based on the data combination results to construct an analysis sample set.

7. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 3, characterized in that, Based on the correlation mining between the rainfall characteristic indicators and the historical slope disaster event records, disaster correlation rules are generated, including: The rainfall characteristic indicators are numerically discretized to determine several level intervals. Based on these level intervals, the rainfall characteristic indicators are transformed to generate independent items. The historical slope disaster event records are treated as special items. Combine the independent items with the special items to construct a transaction item set; The transaction itemset is scanned using an association rule mining algorithm. The first occurrence frequency of each item is counted, and items with the first occurrence frequency not lower than a preset minimum support threshold are selected to form frequent 1-itemsets. Join the frequent 1-itemsets to obtain joined itemsets; prune the joined itemsets to generate candidate 2-itemsets. Scan the transaction itemset to count the second occurrence frequency of the candidate 2-itemsets, filter the candidate 2-itemsets whose second occurrence frequency is not lower than the preset minimum support threshold, and filter out frequent 2-itemsets; Iterate through the above process, generating candidate h-itemsets layer by layer and filtering out frequent h-itemsets until no new frequent itemsets can be generated, and construct disaster association rules.

8. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 1, characterized in that, Rainfall prediction is performed based on the comprehensive database to obtain slope predicted rainfall data; the slope predicted rainfall data is combined with the influence sensitivity coefficient to generate rainfall warning, and a disaster risk index is generated to activate multi-level slope disaster rainfall warning signals for real-time warning of transmission tower slope disaster rainfall, including: Based on the historical and real-time meteorological data accumulated in the comprehensive database, the weather forecasting model is invoked to predict rainfall and obtain slope predicted rainfall data. The predicted rainfall data of the slope is combined with the impact sensitivity coefficient for weighted fusion calculation to generate a disaster risk index; Multiple consecutive early warning threshold intervals are set, each interval corresponding to a risk level; the disaster risk index is compared and determined with the multiple consecutive early warning threshold intervals to generate a target early warning level; Activate multi-level slope disaster rainfall early warning signals based on the target early warning level; The multi-level slope disaster rainfall early warning signals are synchronized to the remote operation and maintenance terminal for disaster rainfall alarm; Among them, the multi-level slope disaster rainfall warning signal includes a standardized response plan corresponding to each level. Different levels of slope disaster rainfall warning signals have different warning names, different response measures, and different information alarm ranges.

9. The method for early warning of rainfall-induced disasters on power transmission tower slopes based on big data analysis according to claim 8, characterized in that, The predicted rainfall data for the slope is combined with the impact sensitivity coefficient for weighted fusion calculation to generate a disaster risk index, including: The slope-predicted rainfall data is matched and calculated with the impact sensitivity coefficient to obtain multiple rainfall impact sensitivity indicators; these multiple rainfall impact sensitivity indicators include cumulative rainfall sensitivity indicators, average rainfall intensity sensitivity indicators, and maximum rainfall intensity sensitivity indicators. Risk calculations are performed based on the cumulative rainfall sensitivity index, the average rainfall intensity sensitivity index, and the maximum rainfall intensity sensitivity index to obtain multiple sub-risk values. The cumulative rainfall sensitive index, the average rainfall intensity sensitive index, and the maximum rainfall intensity sensitive index are weighted and fused according to multiple sub-risk values ​​to obtain multiple weighted risk contribution degrees. A disaster risk index is generated by comprehensively calculating multiple rainfall-sensitive indicators and multiple weighted risk contribution factors.

10. A rainfall-based early warning system for power transmission tower slope disasters based on big data analysis, characterized in that: The system includes: The data acquisition unit is used to collect multidimensional data based on the sloping area of ​​the transmission tower to obtain a multidimensional dataset; and to store the multidimensional dataset in a structured manner to construct a comprehensive database. The rainfall impact index determination unit is used to perform slope disaster association mining on the comprehensive database using big data analysis to generate disaster association rules; and to perform regression calculations based on the disaster association rules to determine rainfall impact indicators. The rainfall sensitivity analysis unit is used to perform rainfall sensitivity analysis on slope disasters based on the rainfall impact index and obtain the impact sensitivity coefficient. The rainfall early warning unit is used to predict rainfall based on the comprehensive database and obtain slope predicted rainfall data; combine the slope predicted rainfall data with the influence sensitivity coefficient to issue a rainfall early warning and generate a disaster risk index to activate a multi-level slope disaster rainfall early warning signal to provide real-time early warning of rainfall disasters on power transmission tower slopes.

Images