Water conservancy project risk early warning analysis method and system based on multi-source information

Abstract

The present invention relates to the field of risk management technology, specifically to a water conservancy project risk early warning analysis method and system based on multi-source information, comprising the following steps: dividing the dam body, flood gate and spillway areas according to the spatial dimension, extracting leakage monitoring records during the flood season, dry season and sudden weather event period, and calculating the trigger factor frequency to generate a trigger factor feature sequence. The present invention, through hierarchical division of spatial and temporal dimensions, combined with the extraction of trigger factor frequency, support and confidence analysis, makes the extraction of risk factor features more comprehensive and accurate. In the construction of a dynamic propagation path, by conducting an in-depth analysis of the spatial trajectory and temporal evolution law of the risk factor, not only the propagation path of the risk factor is clearly displayed, but also the accurate prediction of the risk diffusion area is achieved, providing a basis for the subsequent division of risk levels in the region.

Description

Technical Field

[0001] The present invention relates to the technical field of risk management, and in particular to a water conservancy project risk early warning analysis method and system based on multi-source information. Background Art

[0002] The field of risk management encompasses multiple aspects, including risk identification, risk assessment, and risk control. Its core focus is on reducing or avoiding losses caused by uncertainties through scientific analysis and prediction of potential threats. Information collection and analysis technologies, big data processing technologies, and multi-source information fusion technologies are widely used within risk management to dynamically monitor and provide early warnings for various risks.

[0003] The water conservancy project risk early warning analysis method is a method for analyzing and predicting potential risk factors in water conservancy projects through multi-source information collection, information fusion, and risk assessment. This method covers technical aspects such as the collection of external environmental information, structural characteristics analysis, and real-time monitoring data processing for water conservancy projects. It primarily relies on the collection and fusion of multi-source information, combined with historical operating data for systematic risk modeling and dynamic updating, and uses risk assessment methods to identify potential threats and hidden dangers to water conservancy projects.

[0004] Existing technologies struggle to effectively layer the complex spatial and temporal data of water conservancy projects, resulting in incomplete risk factor extraction and assessment. For example, existing technologies often categorize monitoring data simply along a single dimension, failing to simultaneously incorporate the characteristics of spatial scope and temporal period. This makes it difficult to accurately extract risk factor characteristics across different regions and time periods. Furthermore, existing technologies often rely on static data models to identify risk propagation pathways, making it difficult to dynamically analyze the propagation patterns of risk factors. For example, when leakage monitoring data in a certain area of ​​a dam suddenly increases, existing technologies are unable to track the diffusion trend of this data across the entire region, potentially leading to a lag in the assessment of risks in downstream areas. In the classification and grading of risk factors, existing technologies often rely on a single threshold, making it difficult to handle the ambiguity and uncertainty inherent in complex data. For example, changes in risk levels within a region may be misjudged due to single outliers, impacting the reliability of the overall risk assessment. In terms of time series analysis, existing technologies lack accurate dynamic feature extraction for identifying sudden increases and decreases in risk, making it difficult to capture key changes in risk factors. For example, during the short period of flood season, when leakage data fluctuates violently, existing technologies may fail to identify sudden increases in time and cannot accurately judge the risk change trend in critical time periods. This deficiency may lead to an erroneous assessment of the scope and severity of flood season risk spread, increasing safety hazards of water conservancy projects. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, the embodiment of the present invention provides a water conservancy project risk early warning analysis method and system based on multi-source information. The technical solution is as follows:

[0006] The water conservancy project risk early warning analysis method based on multi-source information includes the following steps:

[0007] S1: Divide the dam, floodgate, and spillway areas according to the spatial dimension, extract leakage monitoring records during the flood season, dry season, and sudden weather event period, calculate the trigger factor frequency, and generate the trigger factor feature sequence;

[0008] S2: extracting the trigger factor frequency in the trigger factor feature sequence, comparing it with the risk threshold to screen the risk factors, analyzing the risk state offset relationship of the segmented intervals, determining the spatial propagation path, and obtaining a dynamic sequence of the risk propagation path;

[0009] S3: Based on the dynamic sequence of the risk propagation path, a fuzzy evaluation index is constructed, the leakage monitoring point data is divided into fuzzy values ​​and weighted, and a fuzzy distribution characteristic of regional risk is generated;

[0010] S4: Perform cluster analysis on the fuzzy distribution characteristics of regional risk, extract the differences in risk factor characteristics of local regions, identify correlations and interactive effects, and form multi-scale risk distribution characteristics after hierarchical classification;

[0011] S5: Extract the time series of risk factors in the multi-scale risk distribution characteristics, predict future change trends, analyze sudden increase points and sudden decrease points to construct a distribution trend curve, and generate water conservancy project risk warning information.

[0012] The present invention has the following improvements: the trigger factor feature sequence includes the support, confidence and trigger frequency of the trigger factor within a time period; the risk propagation path dynamic sequence specifically includes the spatial propagation path, time propagation characteristics and risk state offset relationship of the risk factor; the regional risk fuzzy distribution feature includes the fuzzified risk level, the time series and spatial distribution state corresponding to the risk level; the multi-scale risk distribution feature specifically refers to the risk distribution characteristics within the local area, the correlation characteristics between regions and the hierarchical classification results; the water conservancy project risk warning information includes the sudden increase points, sudden decrease points and corresponding time distribution trend curves of risk changes.

[0013] The present invention has been improved in that the specific steps of dividing the dam body, flood gate and spillway areas according to the spatial dimension, extracting leakage monitoring records during the flood season, dry season and sudden weather event period, and calculating the trigger factor frequency to generate the trigger factor characteristic sequence are as follows:

[0014] S101: Based on the water conservancy project monitoring data, the dam, flood gate, and spillway areas are divided according to the spatial dimension. The data set of each area is divided into the flood season, dry season, and sudden weather event period according to the temporal dimension. The leakage monitoring records of each time period are extracted to generate regional time-divided leakage monitoring records;

[0015] S102: Based on the regional time-divided leakage monitoring records, the leakage monitoring value of each monitoring point is compared with a preset leakage threshold, trigger factors exceeding the leakage threshold are extracted and the trigger frequency is calculated, and statistical analysis is performed based on the trigger frequency to generate a trigger factor frequency analysis result;

[0016] S103: Based on the trigger factor frequency analysis result, according to the time series change trend of the trigger factor, extract the time series characteristics of the trigger factor in each time period, and classify the time series changes to generate a trigger factor feature sequence.

[0017] The present invention has been improved in that the specific steps of extracting the trigger factor frequency in the trigger factor feature sequence, comparing it with the risk threshold to screen the risk factor, analyzing the risk state offset relationship of the segmented interval, and determining the spatial propagation path to obtain the risk propagation path dynamic sequence are as follows:

[0018] S201: Based on the trigger factor feature sequence, the trigger factor frequency in each time period is divided into intervals, the trigger factor frequency in each interval is extracted and compared with a preset risk frequency threshold, risk factors exceeding the risk frequency threshold are screened, and interval risk factor screening results are generated;

[0019] S202: Based on the interval risk factor screening results, analyzing the changes in risk values ​​in the segmented intervals corresponding to the screened risk factors, extracting the state deviation relationship of the risk factors in the segmented intervals, and generating risk state deviation relationship features;

[0020] S203: Based on the risk state offset relationship characteristics, the spatial dimension of the risk factor is tracked, and by analyzing the spatial correlation and time evolution characteristics of the triggering factor and determining the propagation path, a dynamic sequence of the risk propagation path is generated.

[0021] The present invention has been improved in that, based on the dynamic sequence of the risk propagation path, a fuzzy evaluation index is constructed, the leakage monitoring point data is divided into fuzzy values ​​and weighted, and the specific steps of generating the regional risk fuzzy distribution characteristics are as follows:

[0022] S301: Based on the dynamic sequence of risk propagation paths, extract the risk factor propagation paths in the corresponding area, call multi-source monitoring data in the corresponding area, classify and organize the data of the leakage monitoring points, assign risk levels, and generate regional risk level data;

[0023] S302: Based on the regional risk level data, construct a fuzzy evaluation index for the corresponding region, divide the data of each leakage monitoring point into multiple fuzzy values ​​according to the risk level range, and perform fuzzy value archiving processing on the data of all monitoring points to generate regional fuzzy monitoring data;

[0024] S303: Based on the regional fuzzy monitoring data, the time series monitoring data in the region is called, the fuzzy values ​​of the monitoring points in each time period are weighted, the fuzzy distribution state in each time period is obtained, and the regional risk fuzzy distribution characteristics are generated.

[0025] The present invention has the following improvements: for constructing the fuzzy evaluation index of the corresponding area, the formula is adopted:

[0026]

[0027] Calculate the fuzzy value μ(x) of the monitoring point;

[0028] Where x is the actual leakage at the leakage monitoring point, k is the parameter for adjusting the steepness of the fuzzy function curve, c is the median value of the risk level, α is the fuzzy compensation coefficient, and σ is the standard deviation of the regional leakage.

[0029] The present invention is improved in that the fuzzy values ​​of the monitoring points in each time period are weighted and the formula is used:

[0030]

[0031] Calculate the fuzzy distribution state W within the target time period t , a fuzzy value reflecting the overall risk level of the region during the time period, ranging from 0 to 1;

[0032] Among them, w i is the weight of the i-th monitoring point, β i is the time correlation factor of the ith monitoring point, and n is the total number of monitoring points in the time period.

[0033] The present invention has the following improvements: cluster analysis is performed on the regional risk fuzzy distribution characteristics, the differences in risk factor characteristics of local regions are extracted, the correlation and interaction are identified, and the multi-scale risk distribution characteristics are formed after hierarchical classification.

[0034] S401: Based on the regional risk fuzzy distribution characteristics, cluster the risk data of each local area in the distribution characteristics, extract the risk factor difference characteristics between the local areas, and generate local risk factor difference characteristics;

[0035] S402: Based on the local risk factor difference characteristics, identifying the interaction between multiple risk factors, analyzing the distribution characteristics and correlation degree of the risk factors, extracting risk factors whose correlation exceeds a preset correlation threshold, and generating risk factor correlation characteristics;

[0036] S403: Based on the risk factor association characteristics, risk factors exceeding a preset association threshold are graded and classified according to distribution range and characteristics, and the distribution characteristics of the local area and the overall area are combined to form a multi-scale risk distribution characteristic.

[0037] The present invention has the following improvements: extracting the time series of risk factors from the multi-scale risk distribution characteristics, predicting future change trends, analyzing sudden increase and decrease points to construct a distribution trend curve, and generating water conservancy project risk warning information:

[0038] S501: Based on the multi-scale risk distribution characteristics, extract the time series corresponding to each category of risk factors, perform data cleaning and integration on the time series, extract the characteristic values ​​of the risk factor changes in the time series and mark the change trends, and generate the risk factor time series characteristics;

[0039] S502: Based on the risk factor time series characteristics, identify sudden increase points and sudden decrease points of the characteristic values ​​of risk changes in the time series, combine the change rate and characteristic range in the time series, determine the key time period of risk change, and generate risk mutation key time period characteristics;

[0040] S503: Based on the characteristics of the key time period of risk mutation, a distribution trend curve of the time series is constructed. By marking the sudden increase points and sudden decrease points on the time series curve, the visual distribution information of the risk area and time period is extracted to generate water conservancy project risk warning information.

[0041] A water conservancy project risk early warning analysis system based on multi-source information, the system comprising:

[0042] The regional division and feature extraction module divides the dam, floodgate, and spillway areas according to the spatial dimension based on water conservancy project monitoring data. It extracts leakage monitoring records during the flood season, dry season, and sudden weather event period, calculates the trigger factor frequency, and analyzes its changing characteristics in the time dimension to generate a trigger factor feature sequence.

[0043] The risk propagation path analysis module extracts the trigger factor frequency value in each time period based on the trigger factor feature sequence, compares the trigger factor frequency with the risk frequency threshold to screen out the risk factor, analyzes the state deviation characteristics of the risk factor in the segmented interval, determines the propagation path by combining the time and space dimensions, and generates a dynamic sequence of the risk propagation path;

[0044] The fuzzy distribution generation module constructs a fuzzy evaluation index within the region based on the dynamic sequence of the risk propagation path, divides the leakage monitoring point data into fuzzy values, performs weighted processing on the monitoring data according to the fuzzy values, extracts the risk fuzzy distribution of each time period, and generates regional risk fuzzy distribution characteristics;

[0045] The risk factor grading and association module clusters the risk factor characteristics of local areas based on the fuzzy distribution characteristics of regional risks, extracts the difference characteristics between local areas, and classifies the risk factors based on the correlation and interaction between multiple risk factors to generate multi-scale risk distribution characteristics;

[0046] The risk warning and trend prediction module extracts the time series of risk factors based on the multi-scale risk distribution characteristics, analyzes the changing trends in the time series, identifies the sudden increase and decrease points in risk changes, constructs a distribution trend curve based on the trend characteristics, and generates water conservancy project risk warning information.

[0047] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0048] By layering risk factors in spatial and temporal dimensions, combined with trigger factor frequency extraction, support, and confidence analysis, the extraction of risk factor characteristics is more comprehensive and accurate. In constructing dynamic propagation paths, in-depth analysis of the spatial trajectory and temporal evolution of risk factors not only clearly demonstrates the risk factor propagation path but also accurately predicts the risk diffusion area, providing a foundation for subsequent regional risk classification. The introduction of fuzzy evaluation indicators and the fuzzification of monitoring data overcome the uncertainty and ambiguity between different types of monitoring data, enhancing the analysis of risk level distribution characteristics. In clustering regional risk factors, by identifying risk differences and inter-regional correlations, local and global regional risk classification is achieved, establishing a global risk assessment system. Through time series analysis and the identification of sudden increases and decreases, combined with the construction of distribution trend curves, risk changes within key time periods are effectively captured, providing managers with visual dynamic risk warning information. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0050] Figure 1 is a flow chart of the method of the present invention;

[0051] Figure 2 This is a detailed flow chart of step S1 of the present invention;

[0052] Figure 3 This is a schematic diagram of a detailed process of step S2 of the present invention;

[0053] Figure 4 This is a detailed flow chart of step S3 of the present invention;

[0054] Figure 5 This is a detailed flow chart of step S4 of the present invention;

[0055] Figure 6 This is a detailed flow chart of step S5 of the present invention;

[0056] Figure 7 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0057] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0058] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0059] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0060] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0061] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0062] An embodiment of the present invention provides a water conservancy project risk early warning analysis method based on multi-source information, comprising the following steps:

[0063] S1: Divide the dam, floodgate, and spillway areas according to the spatial dimension, extract leakage monitoring records during the flood season, dry season, and sudden weather event period, calculate the trigger factor frequency, and generate the trigger factor feature sequence;

[0064] S2: Extract the trigger factor frequency in the trigger factor feature sequence, compare it with the risk threshold to screen the risk factors, analyze the risk state offset relationship of the segmented interval, determine the spatial propagation path, and obtain the dynamic sequence of the risk propagation path;

[0065] S3: Based on the dynamic sequence of risk propagation paths, a fuzzy evaluation index is constructed, the leakage monitoring point data is divided into fuzzy values ​​and weighted, and the regional risk fuzzy distribution characteristics are generated;

[0066] S4: Cluster analysis is performed on the fuzzy distribution characteristics of regional risks to extract the differences in risk factor characteristics in local areas, identify correlations and interactive effects, and form multi-scale risk distribution characteristics after hierarchical classification;

[0067] S5: Extract the time series of risk factors from the multi-scale risk distribution characteristics, predict future change trends, analyze sudden increase and decrease points to construct distribution trend curves, and generate water conservancy project risk warning information;

[0068] The characteristic sequence of trigger factors includes the support, confidence and trigger frequency of the trigger factors within the time period. The dynamic sequence of risk propagation paths specifically includes the spatial propagation path, temporal propagation characteristics and risk state offset relationship of the risk factors. The fuzzy distribution characteristics of regional risks include the fuzzified risk level, the time series corresponding to the risk level and the spatial distribution state. The multi-scale risk distribution characteristics specifically refer to the risk distribution characteristics within the local area, the correlation characteristics between regions and the grading and classification results. The risk warning information of water conservancy projects includes the sudden increase and decrease points of risk changes and the corresponding time distribution trend curve.

[0069] See also Figure 2 The specific steps for dividing the dam, flood gate, and spillway areas according to the spatial dimension, extracting leakage monitoring records during the flood season, dry season, and sudden weather event period, and calculating the trigger factor frequency to generate the trigger factor characteristic sequence are as follows:

[0070] S101: Based on the water conservancy project monitoring data, the dam, flood gate, and spillway areas are divided according to the spatial dimension. The data set of each area is divided into the flood season, dry season, and sudden weather event period according to the temporal dimension. The leakage monitoring records of each time period are extracted to generate regional time-divided leakage monitoring records;

[0071] Extract leakage-related records based on monitoring equipment deployed in the area, divide the data set of each area into flood season, dry season and sudden weather event period according to the time dimension, conduct statistical analysis on the monitoring data of each time period, and classify and count leakage data. The classification range of leakage includes less than 50m 3 / h, between 50-200m 3 / h and above 200m 3 / h, extract data higher than 200m 3 / h of monitoring data and record the corresponding time period, sort the marked high leakage data according to region and time period, and finally generate regional time-divided leakage monitoring records.

[0072] S102: Based on the regional time-division leakage monitoring records, the leakage monitoring value of each monitoring point is compared with the preset leakage threshold, the trigger factor exceeding the leakage threshold is extracted and the trigger frequency is calculated, and statistical analysis is performed based on the trigger frequency to generate a trigger factor frequency analysis result;

[0073] The leakage monitoring value of each monitoring point is compared with the preset leakage threshold one by one, and the leakage threshold is set to 150m 3 / h, screen out monitoring points that exceed the threshold through logical judgment, and record the corresponding time period and leakage value. Count the number of times each monitoring point exceeds the leakage threshold in each time period, and calculate the trigger frequency of each monitoring point. The trigger frequency is divided into low frequency (less than 3 times per month), medium frequency (3-6 times per month) and high frequency (more than 6 times per month). The upper and lower bounds of the trigger frequency are extracted from historical data and the distribution range is analyzed. The monitoring points that exceed the limit and have a high frequency are marked, and the trigger factor frequency analysis results are generated based on these data.

[0074] S103: Based on the trigger factor frequency analysis results, according to the time series change trend of the trigger factor, extract the time series characteristics of the trigger factor in each time period, and classify the time series changes to generate a trigger factor feature sequence;

[0075] By counting the fluctuation amplitude of the trigger frequency in each time period, the fluctuation amplitude is obtained by calculating the difference between the maximum and minimum values ​​of the frequency, and the fluctuation amplitude is classified into the following ranges: high fluctuation (fluctuation amplitude greater than 50%), medium fluctuation (fluctuation amplitude between 20%-50%) and low fluctuation (fluctuation amplitude less than 20%). The trigger factors in the high fluctuation category are marked as the drastic change category, the trigger factors in the medium fluctuation category are marked as the medium change category, and the trigger factors in the low fluctuation category are marked as the stable change category. The distribution characteristics of the trigger factors of different fluctuation categories in the time series are further analyzed, the distribution trends of the trigger factors of each category in each time period are sorted and classified, the distribution characteristics and change trends of the trigger factors are extracted, and finally the trigger factor feature sequence is generated.

[0076] See also Figure 3 , extract the trigger factor frequency in the trigger factor feature sequence, compare it with the risk threshold to screen the risk factors, analyze the risk state offset relationship of the segmented interval, determine the spatial propagation path, and obtain the specific steps of the risk propagation path dynamic sequence as follows:

[0077] S201: Based on the trigger factor feature sequence, the trigger factor frequency in each time period is divided into intervals, the trigger factor frequency in each interval is extracted and compared with a preset risk frequency threshold, risk factors exceeding the risk frequency threshold are screened, and interval risk factor screening results are generated;

[0078] The process of dividing the trigger factor frequency data in each time period into intervals is to set the interval range to low frequency (less than 3 times), medium frequency (3-6 times) and high frequency (greater than 6 times), extract the trigger factor frequency in each interval, and compare it with the preset risk frequency threshold (the threshold is set to 5 times). By comparing, the risk factors with a frequency higher than the threshold are screened out, the screened high-frequency risk factors are marked and sorted by interval, and finally the interval risk factor screening results are generated.

[0079] S202: Based on the interval risk factor screening results, analyze the changes in the risk values ​​in the segmented intervals corresponding to the screened risk factors, extract the state offset relationship of the risk factors in the segmented intervals, and generate risk state offset relationship features;

[0080] The process of analyzing the changes in risk values ​​of the screened risk factors within the segmented intervals is to extract the risk factor values ​​within each interval, classify the states according to the fluctuations of the risk values ​​within the interval, set the state offset range to significantly increase (the risk value within the interval increases by more than 30% compared with the previous time period), slightly increase (the risk value within the interval increases by between 10%-30%) and decrease (the risk value within the interval decreases by more than 10%), record the state offset relationship of the risk factor in each interval, and classify and organize them according to the offset state, and finally generate the risk state offset relationship characteristics.

[0081] S203: Based on the risk state offset relationship characteristics, track the spatial dimension of the risk factor, analyze the spatial correlation and temporal evolution characteristics of the triggering factor, determine the propagation path, and generate a dynamic sequence of the risk propagation path;

[0082] First, spatial monitoring point data related to the triggering factor was extracted. The spatial correlation criteria were set as follows: the physical distance between monitoring points was less than 1 km, and the risk value change trends within the same time period were similar (the difference in risk value increase or decrease did not exceed 10%). Monitoring points that met these conditions were marked as a correlated monitoring point combination. The average risk value level of the monitoring points in each combination was calculated to represent the risk factor level of the area. For the temporal evolution characteristics, risk factor data for each monitoring point at different time periods was extracted. The temporal evolution judgment range was set as an hourly risk value increase of more than 20%. For monitoring points that met the criteria, the key points of their change (including the starting point and the highest point of the change) were marked in the time series. The direction of change of these monitoring points was further determined. The change direction included a continuous increase, decrease, or stabilization of the risk factor value. Monitoring points with similar change directions were marked as part of the same transmission path. Finally, the propagation trajectory of the risk factor in the spatial dimension was recorded, and the high-risk areas and time periods on the trajectory were marked to organize and generate a dynamic sequence of risk transmission paths.

[0083] See also Figure 4 ,Based on the dynamic sequence of risk propagation path, a fuzzy evaluation index is constructed, the leakage monitoring point data is divided into fuzzy values ​​and weighted, and the specific steps of generating regional risk fuzzy distribution characteristics are as follows:

[0084] S301: Based on the dynamic sequence of risk propagation paths, the risk factor propagation paths within the corresponding area are extracted, multi-source monitoring data for the corresponding area is called, the data of the leakage monitoring points are classified and sorted, and risk levels are assigned to generate regional risk level data;

[0085] Call the multi-source monitoring data in the area, including leakage, soil moisture and surface water level, and classify the data of the monitoring points one by one. According to the actual location of the monitoring points, they are allocated to different regional ranges. The risk level is set as low risk (leakage less than 50m 3 / h, soil moisture below 30%), medium risk (leakage volume between 50-200m 3 / h, soil moisture between 30%-70%) and high risk (leakage greater than 200m 3 / h, soil moisture is higher than 70%), the leakage volume of each monitoring point is extracted and compared with the corresponding preset risk range, the monitoring points with leakage volume in the medium-risk and high-risk ranges are marked, and the risk level of the monitoring point in the current area is analyzed in combination with the historical risk change records of the monitoring point. The data of all monitoring points in the regional range are integrated to finally generate regional risk level data.

[0086] S302: Based on the regional risk level data, construct a fuzzy evaluation index for the corresponding region, divide the data of each leakage monitoring point into multiple fuzzy values ​​according to the risk level range, archive the fuzzy values ​​of the data of all monitoring points, and generate regional fuzzy monitoring data;

[0087] To construct the fuzzy evaluation index of the corresponding area, the formula is used:

[0088]

[0089] Calculate the fuzzy value μ(x) of the monitoring point;

[0090] Where x is the actual leakage at the leakage monitoring point, which is collected by real-time monitoring equipment. For example, a flow meter can directly record the leakage data for each time period. k is a parameter that adjusts the steepness of the fuzzy function curve and determines the transition characteristics of the fuzzy function. The value of k is determined by performing a sensitivity analysis on the distribution range of leakage data in the monitoring area. For example, the final value is determined by comparing the accuracy of curve fitting under different k values.

[0091] c is the middle value of the risk level, which represents the average value of the risk level range. Its value is the mean of the upper and lower bounds of the risk level. For example, the upper and lower bounds of medium risk are 50 and 200m 3 / h, then c is 125m 3 / h, α is the fuzzy compensation coefficient, which reflects the environmental sensitivity or the influence of external disturbances in the region. It is determined by the sensitivity experiment of leakage to risk. For example, the leakage variation under different environmental disturbances (such as rainfall or earthquake conditions) is compared in multiple monitorings, and the fluctuation degree of leakage in the region is analyzed to set α. The higher the fluctuation degree, the larger the value of α. σ is the standard deviation of regional leakage, which reflects the fluctuation degree of leakage data in the region. The standard deviation is calculated by the formula Get, x i is the leakage value of the i-th monitoring point, is the average leakage value of the monitoring points, and n′ is the number of monitoring points.

[0092] Taking the leakage rate of a monitoring point x = 120m3 / h as an example, the risk level range is medium risk (50-200m 3 / h), set c = 125, sensitivity analysis results in k = 0.5, experimental analysis results in fuzzy compensation coefficient α = 0.2, and regional calculation standard deviation σ = 15. Substitute into the formula:

[0093] The fuzzy value μ(x) ranges from 0 to 1 and represents the degree of membership of a monitoring point's leakage risk within a specific risk level. Based on the specific risk level, it can be categorized into the following ranges: Low risk: μ(x) < 0.4 indicates a low risk at the leakage level. The corresponding leakage level at the monitoring point is close to or below the lower limit of the risk level, indicating a high level of safety. Medium risk: 0.4 ≤ μ(x) ≤ 0.7 indicates a moderate risk at the leakage level. The leakage level at the monitoring point is in the middle of the risk level range, falling within the normal monitoring range and requiring attention. High risk: μ(x) > 0.7 indicates a high risk at the leakage level. The leakage level at the monitoring point is close to or exceeds the upper limit of the risk level, potentially posing a threat to project stability and requiring priority attention. The calculated μ(x) is ≈ 0.62, indicating that the leakage risk at the monitoring point falls within the medium risk range. The specific analysis is as follows: Based on the fuzzy value range, μ(x) ≈ 0.62 falls within the medium risk range, indicating that the leakage level at the monitoring point is close to the middle value and within the normal risk level range.

[0094] S303: Based on the regional fuzzy monitoring data, the time series monitoring data in the region is called, the fuzzy values ​​of the monitoring points in each time period are weighted, the fuzzy distribution state in each time period is obtained, and the regional risk fuzzy distribution characteristics are generated;

[0095] The fuzzy values ​​of the monitoring points in each time period are weighted using the formula:

[0096]

[0097] Calculate the fuzzy distribution state W within the target time period t , a fuzzy value reflecting the overall risk level of the region during the time period, ranging from 0 to 1;

[0098] Among them, w i is the weight of the i-th monitoring point. The weight is dynamically adjusted according to the risk level distribution of the area where the monitoring point is located. For example, monitoring points in high-risk areas are given higher weights. The weight value is determined by analyzing the risk level ratio in the area. i is the time correlation factor of the i-th monitoring point, which is used to reflect the impact of the risk change of the monitoring point in the current time period relative to the overall change. The time correlation factor is calculated by the change amplitude of the fuzzy value of the monitoring point in the time period. The formula is Δμ iIt represents the change of the fuzzy value of the i-th monitoring point, n is the total number of monitoring points in the time period, which is directly obtained from the number of monitoring devices. It represents the sum of the changes in the fuzzy values ​​of all monitoring points, which is used to normalize the change in the current monitoring point to calculate the relative impact intensity. j represents the index variable involved in the summation operation, which is used to represent the numbers of all monitoring points, from the 1st monitoring point to the nth monitoring point.

[0099] Taking the five monitoring points in the target area as an example, the weight w i are 1.2, 1.0, 0.8, 1.1 and 1.3 respectively, and the fuzzy value μ of the monitoring point is i (x) are 0.8, 0.5, 0.2, 0.7 and 0.9 respectively, and the time correlation factor β i They are 1.1, 1.0, 0.9, 1.2 and 1.3 respectively. The calculation formula is:

[0100]

[0101] In order to quantify the risk level of the fuzzy distribution state within a time period, W is calculated according to the following range: t The value is divided into: Low risk: W t ≤0.3, indicating that the risk level in the time period is low, the risk fuzzy values ​​of monitoring points in the area are generally small, and the overall state is stable; medium risk: 0.3 <W t ≤0.6, indicating that there is a certain degree of risk within the time period, and the risk fuzzy value of some monitoring points is high, but it has not reached a significant risk state; medium-high risk: 0.6 <W t ≤0.8, indicating that the risk level of the region within the time period is relatively high, and the risk fuzzy values ​​of most monitoring points are high, and attention should be paid to possible development trends; High risk: W t >0.8, indicating that the region as a whole is in a high-risk state during the time period, and the fuzzy values ​​of the monitoring points are highly concentrated in the high-risk range, and emergency management measures need to be taken. According to the division, the calculated result W t ≈0.685 falls into the medium-to-high risk range. t =0.685 indicates that during the current time period, the overall risk level, formed by the superposition of the fuzzy weights of multiple monitoring points within the region, is high, but has not yet reached the highest risk level. This means that although the leakage risk at some monitoring points is significant, it is still within a controllable range.

[0102] See also Figure 5 The specific steps for clustering analysis of regional risk fuzzy distribution characteristics, extracting the differences in local regional risk factor characteristics, identifying correlations and interactive effects, and forming multi-scale risk distribution characteristics after hierarchical classification are as follows:

[0103] S401: Based on the regional risk fuzzy distribution characteristics, cluster the risk data of each local area in the distribution characteristics, extract the risk factor difference characteristics between the local areas, and generate local risk factor difference characteristics;

[0104] The process of clustering the risk data for each local area in the distribution characteristics is as follows: for example, in the dam area of ​​a certain water conservancy project, a total of 50 monitoring points are set up. The fuzzy value of each monitoring point ranges from 0.1 to 0.9. The fuzzy values ​​of some high-risk monitoring points exceed 0.8, and the fuzzy values ​​of monitoring points in other areas are concentrated between 0.3 and 0.6. After normalizing the fuzzy values ​​into a uniform range, a clustering method based on Euclidean distance is used to analyze the risk characteristics between monitoring points. Assuming that the average Euclidean distance between each group of monitoring points is 0.1, after setting the clustering threshold to 0.15, three cluster categories are obtained. The first group of risk points has a fuzzy value above 0.8 and is mainly distributed in the dam discharge area. The second group of risk points has a fuzzy value concentrated between 0.5 and 0.6 and is distributed in the central area of ​​the dam. The third group of risk points has lower fuzzy values ​​and is distributed in the edge area of ​​the dam. Based on the above clustering results, the average risk fuzzy value, the distribution density of risk points and the proportion of high-risk points in different areas are extracted, and the characteristic differences of the three types of clustering results are compared to generate the difference characteristics of local risk factors.

[0105] S402: Based on the differential characteristics of local risk factors, identify the interactive effects between multiple risk factors, analyze the distribution characteristics and correlation levels of the risk factors, extract risk factors whose correlation exceeds a preset correlation threshold, and generate risk factor correlation characteristics;

[0106] The process of identifying the interactions between multiple risk factors involves, for example, performing a time series analysis of the fuzzy values ​​for seepage and water levels in the dam and spillway regions. Using a sliding time window (e.g., set to 1 hour), the change sequence of the fuzzy values ​​for seepage and water levels is extracted and the correlation coefficient between the two is calculated. If, within a certain time period, the fuzzy value for seepage in the dam region increases from 0.6 to 0.8, while the fuzzy value for water levels in the spillway region increases from 0.5 to 0.7, the calculated correlation coefficient is 0.85 (greater than the preset threshold of 0.8), indicating a strong interaction between the two. Further analysis of the directionality of the interaction reveals that when the rate of change in dam seepage exceeds 10%, the rate of change in spillway water levels reaches 15% over the same time period, indicating a causal relationship between the two. By recording these interactions and analyzing their magnitude, the interactions between dam seepage and spillway water levels are identified as key risk factor pairs, generating risk factor correlation signatures.

[0107] S403: Based on the risk factor correlation characteristics, risk factors exceeding the preset correlation threshold are classified according to distribution range and characteristics, and the distribution characteristics of the local area and the overall area are combined to form a multi-scale risk distribution feature;

[0108] Risk factors exceeding pre-set correlation thresholds are classified based on their distribution range and characteristics. For example, in the risk factor correlation features extracted from the dam and spillway regions, seepage fuzzy values ​​and water level fuzzy values ​​are the primary factors. These factors are then classified by calculating their cross-regional coverage ratios and fluctuation amplitudes. Assume that the seepage fuzzy values ​​in the dam region cover 80% of the total monitoring points and have a fluctuation amplitude of 0.3, while the water level fuzzy values ​​in the spillway region cover 60% of the total monitoring points and have a fluctuation amplitude of 0.25. Setting the coverage ratio threshold to 70% and the fluctuation amplitude threshold to 0.2, the seepage fuzzy values ​​in the dam region are classified as high-risk factors, and the water level fuzzy values ​​in the spillway region are classified as medium-risk factors. Furthermore, risk factors with a fluctuation amplitude below 0.1 but a coverage exceeding 90%, such as the meteorological impact fuzzy values ​​in the spillway region, are classified as low-risk factors. Combining these classification results, a cross-regional risk factor classification is generated, and combined with the distribution characteristics of each region, a multi-scale risk distribution feature is formed.

[0109] See also Figure 6 The specific steps for extracting the time series of risk factors from the multi-scale risk distribution characteristics, predicting future change trends, analyzing sudden increase and decrease points to construct distribution trend curves, and generating water conservancy project risk warning information are as follows:

[0110] S501: Based on the multi-scale risk distribution characteristics, extract the time series corresponding to each category of risk factors, clean and organize the time series data, extract the characteristic values ​​of the risk factor changes in the time series and mark the change trends, and generate the risk factor time series characteristics;

[0111] The process of extracting the time series corresponding to each category of risk factors is as follows: for example, in a certain water conservancy project, the time series of high-risk factors in the dam area is extracted as a daily leakage monitoring value series, and the time series of medium-risk factors is extracted as an hourly water level change value series. After filtering out obvious erroneous data points (such as leakage values ​​that are negative or exceed the equipment monitoring range) and duplicate data, the time series is normalized and the time step is unified to the hourly level to ensure that the time series of each risk factor are comparable. In the cleaned data, the characteristic values ​​of the risk factor time series are extracted by calculating the mean, standard deviation and maximum value of the daily leakage value and the water level change, and the risk change characteristics are marked by the change trend of the characteristic value. For example, the leakage value gradually increases (the daily mean value increases from 200m 3 / h increased to 300m 3 / h), while the water level changes show a fluctuating trend (the maximum value fluctuates between 60cm and 80cm), which ultimately generates the time series characteristics of risk factors.

[0112] S502: Based on the time series characteristics of the risk factors, identify the sudden increase and decrease points of the characteristic values ​​of the risk changes in the time series. Combined with the change rate and characteristic range in the time series, determine the key time period of risk change and generate the key time period characteristics of risk mutation;

[0113] The process of identifying sudden increase and decrease points of the characteristic value of risk change in the time series is as follows: for example, in the time series of leakage monitoring values, when the leakage value increases from 300m3 to 100m3 within a certain hour, 3 / h suddenly rises to 450m 3 / h, it is marked as a sudden increase point. In the water level change time series, when the water level drops from 80cm to 50cm and the drop is greater than the set threshold of 30cm, it is marked as a sudden drop point. By combining the change rate of the leakage value time series (for example, the leakage value increases at a rate of 10% per hour within 24 hours) and the characteristic range of water level changes (for example, the maximum water level fluctuates between 60cm and 80cm), the key time periods for risk changes are determined. For example, the three hours before and after the leakage value sudden increase point and the period of time with a large water level drop are identified as key time periods, and ultimately the characteristics of the key time period for risk mutation are generated.

[0114] S503: Based on the characteristics of the key time period of risk mutation, a time series distribution trend curve is constructed. By marking the sudden increase and decrease points on the time series curve, visual distribution information of the risk area and time period is extracted to generate water conservancy project risk warning information;

[0115] The process of constructing the distribution trend curve of the time series is, for example, to draw the time series characteristics of the leakage value and water level change as a broken line trend curve, and mark the sudden increase point and sudden decrease point with red and green dots respectively on the curve. 3 / h suddenly increased to 450m 3 / h of critical time points, and the growth rate within that time period is also noted. On the water level change curve, by continuously marking the locations of sudden drop points, the rapid drop in water level within certain time periods is reflected, and the range and pattern of water level changes are noted through the overall fluctuation trend of the curve. Combined with the above trend curve, the distribution information of each risk factor in the time dimension is further extracted, and the high-risk areas corresponding to the sudden increase in leakage value and the medium-risk areas corresponding to the sudden drop in water level are integrated into one, generating water conservancy project risk warning information with visual characteristics for managers' reference and decision-making.

[0116] See also Figure 7, a water conservancy project risk early warning analysis system based on multi-source information, the system includes:

[0117] The regional division and feature extraction module divides the dam, floodgate, and spillway areas according to the spatial dimension based on water conservancy project monitoring data. It extracts leakage monitoring records during the flood season, dry season, and sudden weather event period, calculates the trigger factor frequency, and analyzes its changing characteristics in the time dimension to generate a trigger factor feature sequence.

[0118] The risk propagation path analysis module extracts the trigger factor frequency value in each time period based on the trigger factor feature sequence, compares the trigger factor frequency with the risk frequency threshold to screen out the risk factors, analyzes the state deviation characteristics of the risk factors in the segmented interval, combines the time and space dimensions to determine the propagation path, and generates a dynamic sequence of risk propagation paths;

[0119] The fuzzy distribution generation module constructs fuzzy evaluation indicators within the region based on the dynamic sequence of risk propagation paths, divides the leakage monitoring point data into fuzzy values, performs weighted processing on the monitoring data according to the fuzzy values, extracts the risk fuzzy distribution of each time period, and generates regional risk fuzzy distribution characteristics;

[0120] The risk factor grading and association module clusters the risk factor characteristics of local areas based on the fuzzy distribution characteristics of regional risks, extracts the difference characteristics between local areas, and combines the correlation and interaction between multiple risk factors to grade the risk factors and generate multi-scale risk distribution characteristics;

[0121] The risk warning and trend prediction module is based on the multi-scale risk distribution characteristics, extracts the time series of risk factors, analyzes the changing trends in the time series, identifies the sudden increase and decrease points in risk changes, constructs the distribution trend curve based on the trend characteristics, and generates water conservancy project risk warning information.

[0122] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0123] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0124] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0125] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0126] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0127] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0128] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0129] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0130] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0131] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A water conservancy project risk early warning analysis method based on multi-source information, characterized by: The following steps are involved: S1: Divide the dam, floodgate, and spillway areas according to the spatial dimension, extract leakage monitoring records during the flood season, dry season, and sudden weather event period, calculate the trigger factor frequency, and generate the trigger factor feature sequence; S2: extracting the trigger factor frequency in the trigger factor feature sequence, comparing it with the risk threshold to screen the risk factors, analyzing the risk state offset relationship of the segmented intervals, determining the spatial propagation path, and obtaining a dynamic sequence of the risk propagation path; S3: Based on the dynamic sequence of the risk propagation path, a fuzzy evaluation index is constructed, and the leakage monitoring point data is divided into fuzzy values ​​and weighted. The specific steps for generating the regional risk fuzzy distribution characteristics are as follows: S301: Based on the dynamic sequence of risk propagation paths, extract the risk factor propagation paths in the corresponding area, call multi-source monitoring data in the corresponding area, classify and organize the data of the leakage monitoring points, assign risk levels, and generate regional risk level data; S302: Based on the regional risk level data, construct a fuzzy evaluation index for the corresponding region, divide the data of each leakage monitoring point into multiple fuzzy values ​​according to the risk level range, and perform fuzzy value archiving processing on the data of all monitoring points to generate regional fuzzy monitoring data; To construct the fuzzy evaluation index of the corresponding area, the formula is used: ； Calculate the fuzzy value of the monitoring point ; in, is the actual leakage amount at the leakage monitoring point, It is a parameter that adjusts the steepness of the fuzzy function curve. is the middle value of risk level, is the fuzzy compensation coefficient, is the standard deviation of regional leakage; S303: Based on the regional fuzzy monitoring data, the time series monitoring data in the region is called, the fuzzy values ​​of the monitoring points in each time period are weighted, the fuzzy distribution state in each time period is obtained, and the regional risk fuzzy distribution characteristics are generated; S4: Perform cluster analysis on the fuzzy distribution characteristics of regional risk, extract the differences in risk factor characteristics of local regions, identify correlations and interactive effects, and form multi-scale risk distribution characteristics after hierarchical classification; S5: Extract the time series of risk factors in the multi-scale risk distribution characteristics, predict future change trends, analyze sudden increase points and sudden decrease points to construct a distribution trend curve, and generate water conservancy project risk warning information.

2. The water conservancy project risk early warning analysis method based on multi-source information according to claim 1 is characterized by: The trigger factor feature sequence includes the support, confidence and trigger frequency of the trigger factor within the time period. The risk propagation path dynamic sequence specifically includes the spatial propagation path, time propagation characteristics and risk state offset relationship of the risk factor. The regional risk fuzzy distribution characteristics include the fuzzified risk level, the time series corresponding to the risk level and the spatial distribution state. The multi-scale risk distribution characteristics specifically refer to the risk distribution characteristics within the local area, the correlation characteristics between regions and the hierarchical classification results. The water conservancy project risk warning information includes the sudden increase points, sudden decrease points of risk changes and the corresponding time distribution trend curve.

3. The water conservancy project risk early warning analysis method based on multi-source information according to claim 1 is characterized by: The specific steps for dividing the dam, flood gate, and spillway areas according to the spatial dimension, extracting leakage monitoring records during the flood season, dry season, and sudden weather event period, and calculating the trigger factor frequency to generate the trigger factor characteristic sequence are as follows: S101: Based on the water conservancy project monitoring data, the dam, flood gate, and spillway areas are divided according to the spatial dimension. The data set of each area is divided into the flood season, dry season, and sudden weather event period according to the temporal dimension. The leakage monitoring records of each time period are extracted to generate regional time-divided leakage monitoring records; S102: Based on the regional time-divided leakage monitoring records, the leakage monitoring value of each monitoring point is compared with a preset leakage threshold, trigger factors exceeding the leakage threshold are extracted and the trigger frequency is calculated, and statistical analysis is performed based on the trigger frequency to generate a trigger factor frequency analysis result; S103: Based on the trigger factor frequency analysis result, according to the time series change trend of the trigger factor, extract the time series characteristics of the trigger factor in each time period, and classify the time series changes to generate a trigger factor feature sequence.

4. The water conservancy project risk early warning analysis method based on multi-source information according to claim 1 is characterized by: The specific steps of extracting the trigger factor frequency in the trigger factor feature sequence, comparing it with the risk threshold to screen the risk factors, analyzing the risk state offset relationship of the segmented interval, and determining the spatial propagation path to obtain the risk propagation path dynamic sequence are as follows: S201: Based on the trigger factor feature sequence, the trigger factor frequency in each time period is divided into intervals, the trigger factor frequency in each interval is extracted and compared with a preset risk frequency threshold, risk factors exceeding the risk frequency threshold are screened, and interval risk factor screening results are generated; S202: Based on the interval risk factor screening results, analyzing the changes in risk values ​​in the segmented intervals corresponding to the screened risk factors, extracting the state deviation relationship of the risk factors in the segmented intervals, and generating risk state deviation relationship features; S203: Based on the risk state offset relationship characteristics, the spatial dimension of the risk factor is tracked, and by analyzing the spatial correlation and time evolution characteristics of the triggering factor and determining the propagation path, a dynamic sequence of the risk propagation path is generated.

5. The water conservancy project risk early warning analysis method based on multi-source information according to claim 1 is characterized by: The fuzzy values ​​of the monitoring points in each time period are weighted using the formula: ； Calculate the fuzzy distribution state within the target time period , a fuzzy value reflecting the overall risk level of the region during the time period, ranging from 0 to 1; in, It is The weight of each monitoring point, It is The time correlation factor of each monitoring point, is the total number of monitoring points in the time period.

6. The water conservancy project risk early warning analysis method based on multi-source information according to claim 1 is characterized by: The specific steps for cluster analysis of the regional risk fuzzy distribution characteristics, extracting the differences in local regional risk factor characteristics, identifying correlations and interactive effects, and forming multi-scale risk distribution characteristics after hierarchical classification are as follows: S401: Based on the regional risk fuzzy distribution characteristics, cluster the risk data of each local area in the distribution characteristics, extract the risk factor difference characteristics between the local areas, and generate local risk factor difference characteristics; S402: Based on the local risk factor difference characteristics, identifying the interaction between multiple risk factors, analyzing the distribution characteristics and correlation degree of the risk factors, extracting risk factors whose correlation exceeds a preset correlation threshold, and generating risk factor correlation characteristics; S403: Based on the risk factor association characteristics, risk factors exceeding a preset association threshold are graded and classified according to distribution range and characteristics, and the distribution characteristics of the local area and the overall area are combined to form a multi-scale risk distribution characteristic.

7. The water conservancy project risk early warning analysis method based on multi-source information according to claim 1 is characterized by: The specific steps of extracting the time series of risk factors from the multi-scale risk distribution characteristics, predicting future change trends, analyzing sudden increase and decrease points to construct a distribution trend curve, and generating water conservancy project risk warning information are as follows: S501: Based on the multi-scale risk distribution characteristics, extract the time series corresponding to each category of risk factors, perform data cleaning and integration on the time series, extract the characteristic values ​​of the risk factor changes in the time series and mark the change trends, and generate the risk factor time series characteristics; S502: Based on the risk factor time series characteristics, identify sudden increase points and sudden decrease points of the characteristic values ​​of risk changes in the time series, combine the change rate and characteristic range in the time series, determine the key time period of risk change, and generate risk mutation key time period characteristics; S503: Based on the characteristics of the key time period of risk mutation, a distribution trend curve of the time series is constructed. By marking the sudden increase points and sudden decrease points on the time series curve, the visual distribution information of the risk area and time period is extracted to generate water conservancy project risk warning information.

8. The water conservancy project risk early warning analysis system based on multi-source information is characterized by: The method for water conservancy project risk early warning analysis based on multi-source information according to any one of claims 1 to 7 is implemented, wherein the system comprises: The regional division and feature extraction module divides the dam, floodgate, and spillway areas according to the spatial dimension based on water conservancy project monitoring data. It extracts leakage monitoring records during the flood season, dry season, and sudden weather event period, calculates the trigger factor frequency, and analyzes its changing characteristics in the time dimension to generate a trigger factor feature sequence. The risk propagation path analysis module extracts the trigger factor frequency value in each time period based on the trigger factor feature sequence, compares the trigger factor frequency with the risk frequency threshold to screen out the risk factor, analyzes the state deviation characteristics of the risk factor in the segmented interval, determines the propagation path by combining the time and space dimensions, and generates a dynamic sequence of the risk propagation path; The fuzzy distribution generation module constructs a fuzzy evaluation index within the region based on the dynamic sequence of the risk propagation path, divides the leakage monitoring point data into fuzzy values, performs weighted processing on the monitoring data according to the fuzzy values, extracts the risk fuzzy distribution of each time period, and generates regional risk fuzzy distribution characteristics; The risk factor grading and association module clusters the risk factor characteristics of local areas based on the fuzzy distribution characteristics of regional risks, extracts the difference characteristics between local areas, and classifies the risk factors based on the correlation and interaction between multiple risk factors to generate multi-scale risk distribution characteristics; The risk warning and trend prediction module extracts the time series of risk factors based on the multi-scale risk distribution characteristics, analyzes the changing trends in the time series, identifies the sudden increase and decrease points in risk changes, constructs a distribution trend curve based on the trend characteristics, and generates water conservancy project risk warning information.

Images