Earth and rockfill dam seepage safety monitoring, forecasting and early warning system

Abstract

The invention provides an earth and rockfill dam seepage safety monitoring, forecasting and early warning system, and relates to the technical field of earth and rockfill dam seepage safety early warning, the earth and rockfill dam seepage safety monitoring, forecasting and early warning system comprises a data acquisition module, a data processing and analysis module, a safety early warning analysis module, a safety early warning pushing module and a safety early warning visualization module, analyzing the collected monitoring data to obtain a scheme, and then determining a monitoring index; the safety early warning analysis module is used for constructing a safety early warning model and drawing up an early warning response coefficient; according to the technical key points, a data acquisition module, a data processing analysis module, a safety early warning analysis module, a safety early warning push module and a safety early warning visualization module are integrated, and a seepage monitoring index drawing model of an improved POT model parameter threshold selection method is established based on a chaotic mapping optimization algorithm; on the basis of real-time analysis and early warning of a seepage monitoring safety state, the accuracy of seepage monitoring, forecasting and early warning under a complex seepage condition can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of earth-rock dam seepage safety early warning technology, in particular to an earth-rock dam seepage safety monitoring, forecasting and early warning system. Background Art

[0002] An earth-rock dam generally refers to a retaining dam constructed from locally sourced earth, stone, or a mixture of these materials through dumping, filling, and compaction. When the dam's body is primarily composed of earth and gravel, it's called an earth dam; when it's primarily composed of slag, pebbles, or blasted stone, it's called a rockfill dam. When both types of local materials make up a significant proportion, it's called an earth-rock dam. Earth-rock dams are the oldest type of dam.

[0003] Earth-rock dams, characterized by safety, economy, and adaptability, have become a globally recognized economical and adaptable dam type. Seepage is a key factor affecting the safety of earth-rock dams, and various forms of failure in earth-rock dams are directly or indirectly related to seepage. Therefore, safety monitoring of earth-rock dams is essential. Monitoring indicators define the safety limits of effect quantities and provide a scientific basis for determining whether structural performance is normal. They can help managers develop safe operation and maintenance plans and promptly identify potential safety hazards.

[0004] The various schemes and systems for formulating the existing earth-rock dam seepage monitoring indicators have the following technical problems:

[0005] (1) When using the typical small probability method for earth-rock dam seepage early warning, it is limited by the fact that it generally only analyzes the extreme values ​​in the monitoring sequence. Therefore, in actual application, it relies on historical data or standards to set fixed thresholds, which cannot reflect the dynamic evolution of earth-rock dam seepage and is prone to omissions or false alarms.

[0006] (2) In the process of formulating threshold parameters using the POT model, the commonly used methods of the mean function graph method and the Hill graph method are to formulate thresholds by subjectively analyzing the trend of the graph. However, they are difficult to program and have large errors. The sample kurtosis selection method has a clear theoretical basis, but its calculation process is complicated and is not suitable for systematic integration into the program.

[0007] (3) By constructing a threshold increasing sequence and calculating the corresponding monitoring indicators under different threshold conditions, the improved threshold determination method combined with the 3σ criterion in probability theory is proposed. However, the intelligent optimization algorithm has the problem of being prone to local optimality and low precision in the optimal selection of thresholds, and has a small search range and weak search ability.

[0008] (4) The monitoring and early warning management of seepage in existing earth-rock dams mainly focuses on the collection, storage, and retrieval of monitoring data, as well as the monitoring and early warning visualization platform. Currently, the commonly used solutions have problems such as low accuracy, insufficient response forecasting and early warning capabilities under complex conditions, and low intelligence, making it difficult to achieve real-time dam safety monitoring and early warning. Summary of the Invention

[0009] In order to overcome the shortcomings of the existing reliance on manual experience to set fixed warning thresholds, which is difficult to adapt to complex geological conditions and dynamic environmental changes, resulting in insufficient warning accuracy, the embodiment of the present application provides a seepage safety monitoring, forecasting and warning system for earth-rock dams. By integrating a data acquisition module, a data processing and analysis module, a safety warning analysis module, a safety warning push module and a safety warning visualization module, and based on an improved optimization algorithm, a seepage monitoring index formulation model of an improved POT model parameter threshold selection method is established. On the basis of real-time analysis and warning of the seepage monitoring safety status, the accuracy of seepage monitoring and forecasting under complex seepage conditions can be improved.

[0010] The technical solution adopted by the embodiment of the present application to solve the technical problem is:

[0011] A seepage safety monitoring, forecasting and early warning system for earth-rock dams, comprising a data acquisition module, a data processing and analysis module, a safety early warning analysis module, a safety early warning push module and a safety early warning visualization module, all of which are integrated into the system;

[0012] The data acquisition module is used to collect historical and real-time monitoring raw data, including reservoir water level, rainfall and dam seepage monitoring data, and integrate them into a data sequence table to provide a data basis for subsequent analysis and early warning;

[0013] The data processing and analysis module analyzes the collected monitoring data based on the intelligent optimization algorithm and the POT model, and determines the monitoring indicators after obtaining the solution;

[0014] The safety early warning analysis module builds a safety early warning model based on the monitoring indicator analysis results of the data processing and analysis module and combines the status information of the foundation parameters of the earth-rock dam, formulates the early warning response coefficient, sets the corresponding safety early warning level, and conducts early warning analysis;

[0015] The security warning push module activates the corresponding emergency response mechanism according to the security warning level determined by the security warning analysis module, sends alarm information to relevant personnel, and activates the emergency plan at the same time;

[0016] The security warning visualization module is used to present monitoring data, analysis results, and warning alarm information in a visual form and generate reports.

[0017] Based on the above earth-rock dam seepage safety monitoring, forecasting and early warning system, the following rock dam seepage safety monitoring and forecasting model method is used to work:

[0018] S1: Obtain data such as water level of earth-rock dam reservoir and historical monitoring of water head at monitoring points;

[0019] S2: Organize the data and perform preprocessing operations. Remove duplicate, erroneous, and missing data through data cleaning to ensure data quality. Convert data from different sources and formats into a unified format through data conversion.

[0020] S3: Analyze the pre-processed monitoring data, extract characteristic indicators related to safety hazards and abnormal patterns, build an abnormality prediction model, obtain abnormal condition indicators, and clarify the correlation between real-time monitoring status and potential safety hazards and abnormal patterns;

[0021] Among them, the constructed abnormal prediction model uses the threshold parameters in the indicator formulation model to analyze the characteristics of the tail data of the random sequence of monitoring data, and the threshold parameter determination method is determined by the following expression:

[0022] Conditional distribution function F of excess sequence T (y) is:

[0023] F T (y)=P(xT≤y|x>T)

[0024] The expression of F(x) with respect to F(y) is:

[0025] F(x)=F T (y)[1-F(T)]+F(T)

[0026] Among them, F(x) is the monitoring index x under the proposed significance level α α The basis of x α =F -1 (x, α)

[0027] There is a correlation between the distribution functions of the original measurement sequence, the over-threshold measurement sequence and the excess sequence, and a relationship can be constructed to solve it.

[0028] The PBdH theorem in extreme value theory shows that for a sufficiently large threshold T, the conditional distribution function F of the excess y T (y) converges to the generalized Pareto distribution, that is:

[0029]

[0030] Among them, ξ T ,σ T are two evaluation parameters in the POT model;

[0031] S4: Based on the abnormal condition indicators proposed in S3, predict the seepage safety status.

[0032] In one possible implementation, during the calculation of the threshold T, the Logistic-Tent chaotic mapping optimization algorithm is used to update the data, and its expression is:

[0033]

[0034] Among them, r∈(0,4];

[0035] The finder position update equation is:

[0036]

[0037] The calculation formula of parameter ω is:

[0038]

[0039] Where, ω0 is a given positive real number; t is the current iteration number; t0 is the given iteration number;

[0040] The sparrow position update formula based on Levy flight is:

[0041]

[0042] Where γ is the step control parameter; Levy(λ) satisfies Levy~u=t -λ 1<λ<3;

[0043] The sparrow position update based on reverse learning is:

[0044]

[0045] The dynamic selection strategy method is: when rand∈(0, 0.5], select the Levy flight strategy to update the sparrow's position. Otherwise, select the reverse learning strategy to update the sparrow's position.

[0046] In one possible implementation, during the process of earth-rock dam seepage safety monitoring and early warning, multiple safety warning levels are set. The method for setting multiple safety warning levels is as follows:

[0047] S1. Divide the levels of monitoring indicators according to the 3σ criterion in probability theory. The above definition can be expressed as:

[0048] x 4.5％ =E(x)±2σ,x 0.3％ =E(x)±3σ

[0049] Among them, E(x) and σ are the mathematical expectation and standard deviation of the effect size respectively; x 4.5％ 、x 0.3％ The significance level x 4.5％ and x 0.3％ The monitoring indicator value at that time, that is, the warning value and danger value of the monitoring indicator.

[0050] S2: The correlation between the measured value sequence and the proposed monitoring indicators is:

[0051] x 0.3％ -x 4.5％ =±σ

[0052] S3: For the original measurement sequence {x i}Construct a threshold random sequence {T1, ..., T j ,…,T N}. For each threshold T j , construct the corresponding over-threshold measurement sequence and excess sequence, and calculate the corresponding monitoring indicator warning value x 4.5％ T j and the risk value x 0.3％ T j .

[0053] S4: In the threshold random sequence, the most reasonable threshold T best satisfy:

[0054] c j =|Δ j -σ|→0

[0055] Among them, Δ j is x 4.5％ T j with x 0.3％ T j The difference, Δ j =x 0.3％ T j -x 4.5％ T j ;c j is Δ j The absolute value of the difference from the standard deviation σ of the original monitoring sequence, c j =|Δ j -σ|.

[0056] In a possible implementation, the seepage monitoring warning indicators in S1 include warning value and danger value warning indicators, which correspond to confidence levels α=4.5% and 0.3% respectively.

[0057] In a possible implementation, the safety warning visualization module for earth-rock dam seepage safety detection and warning includes a display layer (user interaction interface), a data processing layer (core functional module) and a storage layer (data management).

[0058] Among them, the display layer (user interaction interface) realizes user login, data visualization, comprehensive early warning and reporting system and information management; the data processing layer (core functional module) is used to formulate early warning indicators, chart generation and comprehensive early warning; the storage layer is used to manage historical data and real-time data collection.

[0059] In one possible implementation, the security warning visualization module supports data flow and function linkage. After the user logs in, the real-time / historical seepage data and warning status can be viewed through the data visualization interface; the data processing layer calls data from the storage layer, completes the dynamic calculation of warning indicators and chart generation, and feeds the results back to the display layer, and realizes global risk integration through the comprehensive warning module.

[0060] The beneficial effects of this application are:

[0061] First, in this scheme, by integrating the data acquisition module, data processing and analysis module, safety warning analysis module, safety warning push module and safety warning visualization module, and based on the chaos mapping optimization algorithm, an improved POT model parameter threshold selection method is established to formulate a seepage monitoring indicator model. This can improve the accuracy of seepage monitoring forecast under complex seepage conditions on the basis of real-time analysis and early warning of seepage monitoring safety status;

[0062] Second, in this solution, by integrating the display layer, data processing layer and storage layer on the safety warning visualization module to continuously update the monitoring data and warning indicators, real-time warning of dam seepage safety can be achieved, and the dam safety status can be evaluated in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a structural diagram of a seepage safety monitoring and early warning system for earth-rock dams according to the present invention;

[0064] Figure 2 This is a schematic diagram of the structure of an improved intelligent optimization algorithm of the present invention;

[0065] Figure 3 This is a flow chart of a method for formulating early warning indicators for earth-rock dam seepage safety monitoring according to the present invention;

[0066] Figure 4 This is a structural diagram of the earth-rock dam seepage safety monitoring and early warning visualization platform of the present invention. DETAILED DESCRIPTION

[0067] The technical solution in the embodiments of the present application is to solve the problems of the above-mentioned background technology, and the overall idea is as follows:

[0068] This application introduces the specific structure of a seepage safety monitoring, forecasting and early warning system for earth-rock dams. Figure 1-Figure 4 As shown, the system includes a data acquisition module, a data processing and analysis module, a security warning analysis module, a security warning push module and a security warning visualization module, all of which are integrated into the system;

[0069] The data acquisition module is used to collect historical and real-time monitoring raw data, including reservoir water level, rainfall and dam seepage monitoring data, and integrate them into a data sequence table to provide a data basis for subsequent analysis and early warning, ensuring the accuracy and real-time nature of the data.

[0070] The data processing and analysis module is based on the intelligent optimization algorithm and the POT model (the POT model believes that the original measurement sequence is a measurement sequence of an independent random variable x with a distribution function F(x) {x1, x2, ..., x N}; Select T within the upper and lower limits of the measurement sequence, which will satisfy x j The variable of >T is called super-threshold measurement, which is constructed as super-threshold measurement sequence; j =x j The value of -T is called excess, which is constructed as the excess sequence {y j The POT model introduces the concept of thresholds, focusing on analyzing the characteristics of the tail data of a random sequence. By analyzing the collected monitoring data and determining monitoring indicators after obtaining a plan, the prediction accuracy of the early warning system can be improved.

[0071] The safety early warning analysis module builds a safety early warning model based on the monitoring indicator analysis results of the data processing and analysis module and combines the status information of the earth-rock dam foundation parameters. It also formulates the early warning response coefficient and sets the corresponding safety early warning level to conduct early warning analysis, which can ensure the accuracy and timeliness of early warning alarms and reduce false alarms and missed alarms.

[0072] The safety warning push module activates the corresponding emergency response mechanism based on the safety warning level determined by the safety warning analysis module, and sends alarm information to relevant personnel. It quickly notifies relevant personnel through multiple channels to ensure the timeliness and coverage of information transmission, and at the same time activates the emergency plan to reduce accident losses;

[0073] The safety warning visualization module is used to display monitoring data, analysis results, and warning alarm information in a visual form and generate various reports, thus providing an intuitive and clear graphical interface to facilitate managers and technicians to quickly understand and handle dam conditions;

[0074] Secondly, when formulating monitoring indicators, the data processing and analysis module extracts time series monitoring data and performs preprocessing operations. Data cleaning removes duplicate, erroneous, and missing data to ensure data quality. Data conversion converts data from different sources and formats into a unified format to facilitate subsequent processing. Data normalization ensures that all features are on the same scale.

[0075] Next, the pre-processed monitoring data is analyzed to extract characteristic indicators related to safety hazards and abnormal patterns, build a construction anomaly prediction model, obtain abnormal condition indicators, and clarify the correlation between real-time monitoring status and potential safety hazards and abnormal patterns;

[0076] like Figure 4 As shown in the figure, the safety warning visualization module for earth-rock dam seepage safety detection and warning includes a display layer (user interaction interface), a data processing layer (core functional module) and a storage layer (data management);

[0077] The display layer (user interface) implements user login (provides identity authentication to ensure the security of data access and operation permissions), data visualization (dynamically displays seepage monitoring data and supports multiple views such as charts and maps), comprehensive early warning (real-time display of seepage safety warning levels (such as blue / yellow / orange / red) and risk locations), and reporting system and information management (generates monitoring reports and supports data export and historical record management functions). The data processing layer (core functional module) is used to formulate early warning indicators.

[0078] Among them, the data processing layer (core functional module): formulates early warning indicators: dynamically optimizes seepage thresholds (such as seepage pressure and seepage volume critical values) based on intelligent algorithms, generates charts (automatically draws seepage parameter time curves, bar charts, etc., to intuitively reflect data change trends) and comprehensive early warning (integrates multi-source data (water level, rainfall) and early warning indicators to trigger graded early warning signals); the storage layer is used to manage historical data (archive long-term monitoring data, support trend analysis and model training) and real-time data collection (continuously receives and updates current seepage parameters through the sensor network to ensure the timeliness of early warnings);

[0079] Secondly, the security warning visualization module supports data flow and functional linkage. After the user logs in, they can view real-time / historical infiltration data and warning status through the data visualization interface. The data processing layer calls data from the storage layer to complete the dynamic calculation of warning indicators and generate charts. The results are fed back to the display layer, and global risk integration is achieved through the comprehensive warning module.

[0080] Based on the above earth-rock dam seepage safety monitoring, forecasting and early warning system, the following rock dam seepage safety monitoring and forecasting model method is used to work:

[0081] S1: Obtain data such as water level of earth-rock dam reservoir and historical monitoring of water head at monitoring points;

[0082] Among them, the seepage monitoring early warning indicators include warning value and danger value early warning indicators, corresponding to confidence levels α = 4.5% and 0.3% respectively;

[0083] S2: Organize the data and perform preprocessing operations. Remove duplicate, erroneous, and missing data through data cleaning to ensure data quality. Convert data from different sources and formats into a unified format through data conversion.

[0084] S3: Analyze the pre-processed monitoring data, extract characteristic indicators related to safety hazards and abnormal patterns, build an abnormality prediction model, obtain abnormal condition indicators, and clarify the correlation between real-time monitoring status and potential safety hazards and abnormal patterns.

[0085] S4: Based on the abnormal condition indicators proposed in S3, predict the seepage safety status.

[0086] In one possible implementation, the anomaly prediction model constructed by S3 uses the threshold parameters in the indicator formulation model to analyze the characteristics of the tail data of the random sequence of monitoring data, and the threshold parameter determination method is determined by the following expression:

[0087] Conditional distribution function F of excess sequence T (y) is:

[0088] F T (y)=P(xT≤y|x>T)

[0089] The expression of F(x) with respect to F(y) is:

[0090] F(x)=F T (y)[1-F(T)]+F(T)

[0091] Among them, F(x) is the monitoring index x under the proposed significance level α α The basis of x α =F -1 (x, α)

[0092] There is a correlation between the distribution functions of the original measurement sequence, the over-threshold measurement sequence and the excess sequence, and a relationship can be constructed to solve it.

[0093] For a sufficiently large threshold T, the conditional distribution function F of the excess y T (y) converges to the generalized Pareto distribution, that is:

[0094]

[0095] Among them, ξT , σ T are two evaluation parameters in the POT model;

[0096] According to the PBdH theorem, by setting the threshold T, the original measurement sequence {x i} to construct the excess sequence {y j} and obtain its distribution function F T (y). The distribution function F corresponding to the excess sequence T (y) has a corresponding relationship with the distribution function F(x) corresponding to the original measurement value sequence, so the distribution function F(x) of the corresponding original measurement value sequence can be solved for any set threshold T that meets the conditions;

[0097] Therefore, when the distribution function F(x) of the original measurement sequence is obtained, the monitoring indicator x under the proposed significance level α can be determined. α In summary, we get the threshold T and monitoring index x α The inherent correlation between them can be used to formulate reasonable monitoring indicators;

[0098] In the process of calculating the threshold T, the Logistic-Tent chaotic mapping optimization algorithm (intelligent optimization algorithm) is used to update the data, and its expression is:

[0099]

[0100] Among them, r∈(0,4];

[0101] In the early iterations, the discoverer uses a larger inertia weight to conduct global exploration at a larger scale and quickly find the global optimal solution. In the later iterations, the discoverer uses a smaller inertia weight to improve local development capabilities and accelerate convergence, while avoiding the problem of falling into the local optimal solution. The discoverer's position update equation is:

[0102]

[0103] Among them, x i，j is the j-th dimension position of the i-th sparrow in the population, N is the maximum number of iterations set by the initial conditions, t is the current number of iterations, α1, α2 and R are random numbers that obey a uniform distribution between (0, 1], c is a random number that obeys a normal distribution, and ST is the safety index value set by the initial conditions;

[0104] The calculation formula of parameter ω is:

[0105]

[0106] Where, ω0 is a given positive real number; t is the current iteration number; t0 is the given iteration number;

[0107] The sparrow position update formula based on Levy flight is:

[0108]

[0109] Where γ is the step control parameter; Levy(λ) satisfies Levy~u=t -λ 1<λ<3;

[0110] The sparrow position update based on reverse learning is:

[0111]

[0112] Based on the two aforementioned methods, a dynamic selection strategy is used to update the sparrow's position, alternating between Levy flight and reverse learning with a certain probability. In the Levy flight strategy, a step factor is used to expand the search range and escape the local optimum. Simultaneously, the reverse learning strategy increases the diversity of solutions, improving the algorithm's search optimization performance.

[0113] The dynamic selection strategy method is: when rand∈(0, 0.5], select the Levy flight strategy to update the sparrow's position. Otherwise, select the reverse learning strategy to update the sparrow's position.

[0114] In some examples, engineering applications use the 3σ criterion in probability theory to classify monitoring indicators. The effect size in a random distribution sequence has a probability of about 4.5% falling outside (μ-2σ, μ+2σ), which can be regarded as a low-probability event and used as a "warning value x 4.5％ "Based on the principle that the surface monitoring object changes from normal working state to abnormal working state; the effect size has a probability of about 0.3% falling outside (μ-3σ, μ+3σ), which can be regarded as an impossible event as the "hazard value x 0.3％ " is based on the fact that the monitored object changes from abnormal working state to dangerous working state;

[0115] Based on the above characteristics, multiple safety warning levels are set for earth-rock dam seepage safety monitoring and early warning. The method for formulating multiple safety warning levels is as follows:

[0116] S1. Divide the levels of monitoring indicators according to the 3σ criterion in probability theory. The above definition can be expressed as:

[0117] x 4.5％ =E(x)±2σ,x 0.3％ =E(x)±3σ

[0118] Among them, E(x) and σ are the mathematical expectation and standard deviation of the effect size respectively; x 4.5％ 、x 0.3％ The significance level x 4.5％and x 0.3％ The monitoring indicator value at that time, that is, the warning value and danger value of the monitoring indicator;

[0119] S2: The correlation between the measured value sequence and the proposed monitoring indicators is:

[0120] x 0.3％ -x 4.5％ =±σ

[0121] S3: For the original measurement sequence {x i}Construct a threshold random sequence {T1, ..., T j ,…,T N}. For each threshold T j , construct the corresponding over-threshold measurement sequence and excess sequence, and calculate the corresponding monitoring indicator warning value x 4.5％ T j and the risk value x 0.3％ T j .

[0122] S4: In the threshold random sequence, the most reasonable threshold T best satisfy:

[0123] c j =|Δ j -σ|→0

[0124] Among them, Δ j is x 4.5％ T j with x 0.3％ T j The difference, Δ j =x 0.3％ T j -x 4.5％ T j ;c j is Δ j The absolute value of the difference from the standard deviation σ of the original monitoring sequence, c j =|Δ j -σ|.

[0125] like Figure 3 As shown, extract time series monitoring data and perform preprocessing operations. Through data cleaning, remove duplicate, erroneous, and missing data to ensure data quality. Through data conversion, convert data from different sources and formats into a unified format to facilitate subsequent processing. By normalizing the data, ensure that all features are on the same scale;

[0126] By analyzing the pre-processed monitoring data, we can extract characteristic indicators related to safety hazards and abnormal patterns, build a construction anomaly prediction model, obtain abnormal condition indicators, and clarify the correlation between real-time monitoring status and potential safety hazards and abnormal patterns.

[0127] Finally, it should be noted that the above embodiments are merely examples for the purpose of illustrating the present invention and are not intended to limit the embodiments. Those skilled in the art will readily appreciate that other variations or modifications based on the above description are possible. It is not necessary and impossible to provide an exhaustive list of all embodiments. However, obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A seepage safety monitoring, forecasting and early warning system for earth-rock dams, characterized in that: Including data acquisition module, data processing and analysis module, safety warning analysis module, safety warning push module and safety warning visualization module all integrated into the system; The data acquisition module is used to collect historical and real-time monitoring raw data, including reservoir water level, rainfall and dam seepage monitoring data, and integrate them into a data sequence table to provide a data basis for subsequent analysis and early warning; The data processing and analysis module analyzes the collected monitoring data based on the intelligent optimization algorithm and the POT model, and determines the monitoring indicators after obtaining the solution; The safety early warning analysis module builds a safety early warning model based on the monitoring indicator analysis results of the data processing and analysis module and combines the status information of the foundation parameters of the earth-rock dam, formulates the early warning response coefficient, sets the corresponding safety early warning level, and conducts early warning analysis; The security warning push module activates the corresponding emergency response mechanism according to the security warning level determined by the security warning analysis module, sends alarm information to relevant personnel, and activates the emergency plan at the same time; The security warning visualization module is used to present monitoring data, analysis results, and warning alarm information in a visual form and generate reports.

2. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 1, characterized in that: Based on the above earth-rock dam seepage safety monitoring, forecasting and early warning system, the following rock dam seepage safety monitoring and forecasting model method is used: S1: Obtain data such as water level of earth-rock dam reservoir and historical monitoring of water head at monitoring points; S2: Organize the data and perform preprocessing operations. Remove duplicate, erroneous, and missing data through data cleaning to ensure data quality. Convert data from different sources and formats into a unified format through data conversion. S3: Analyze the pre-processed monitoring data, extract characteristic indicators related to safety hazards and abnormal patterns, build an abnormality prediction model, obtain abnormal condition indicators, and clarify the correlation between real-time monitoring status and potential safety hazards and abnormal patterns. S4: Based on the abnormal condition indicators proposed in S3, predict the seepage safety status.

3. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 2, characterized in that: The anomaly prediction model constructed by S3 adopts the threshold parameters in the indicator formulation model to analyze the characteristics of the tail data of the random sequence of monitoring data, and the threshold parameter determination method is determined by the following expression: Conditional distribution function F of excess sequence T (y) is: F T (y)=P(x-T≤y|x>T) The expression of F(x) with respect to F(y) is: F(x)=F T (y)[1-F(T)]+F(T) Among them, F(x) is the monitoring index x under the proposed significance level α α The basis of x α =F -1 (x, α) There is a correlation between the distribution functions of the original measurement sequence, the over-threshold measurement sequence and the excess sequence, and a relationship can be constructed to solve it. For a sufficiently large threshold T, the conditional distribution function F of the excess y T (y) converges to the generalized Pareto distribution, that is: Among them, ξ T ,σ T are two evaluation parameters in the POT model.

4. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 3, characterized in that: In the process of calculating the threshold T, the Logistic-Tent chaotic mapping and dynamic selection strategy are used to improve the optimization algorithm, and its expression is: Among them, r∈(0,4]; The discoverer position update equation is: The calculation formula of parameter ω is: Where, ω0 is a given positive real number; t is the current iteration number; t0 is the given iteration number; The sparrow position update formula based on Levy flight is: Where γ is the step control parameter; Levy(λ) satisfies Levy~u=t -λ 1<λ<3; The sparrow position update based on reverse learning is: The dynamic selection strategy method is: when rand∈(0, 0.5], select the Levy flight strategy to update the sparrow's position. Otherwise, select the reverse learning strategy to update the sparrow's position.

5. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 1, characterized in that: In the process of earth-rock dam seepage safety monitoring and early warning, multiple safety warning levels are set for it. The method of formulating multiple safety warning levels is as follows: S1. Divide the levels of monitoring indicators according to the 3σ criterion in probability theory. The above definition can be expressed as: x 4.5％ =E(x)±2σ,x 0.3％ =E(x)±3σ Among them, E(x) and σ are the mathematical expectation and standard deviation of the effect size respectively; x 4.5％ 、x 0.3％ The significance level x 4.5％ and x 0.3％ The monitoring indicator value at that time, that is, the warning value and danger value of the monitoring indicator. S2: The correlation between the measured value sequence and the proposed monitoring indicators is: x 0.3％ -x 4.5％ =±σ S3: For the original measurement sequence {x i }Construct a threshold random sequence {T1, ..., T j ,…,T N }. For each threshold T j , construct the corresponding over-threshold measurement sequence and excess sequence, and calculate the corresponding monitoring indicator warning value x 4.5％ T j and the risk value x 0.3％ T j . S4: In the threshold random sequence, the most reasonable threshold T best satisfy: c j =|D j -σ|→0 Among them, Δ j is x 4.5％ T j with x 0.3％ T j The difference, Δ j =x 0.3％ T j -x 4.5％ T j ;c j is Δ j The absolute value of the difference from the standard deviation σ of the original monitoring sequence, c j =|Δ j -σ|.

6. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 5, characterized in that: The seepage monitoring early warning indicators in S1 include warning value and danger value early warning indicators, which correspond to confidence levels α=4.5% and 0.3% respectively.

7. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 1, characterized in that: The safety warning visualization module for earth-rock dam seepage safety detection and warning includes a display layer (user interaction interface), a data processing layer (core functional module) and a storage layer (data management). The presentation layer (user interaction interface) implements user login, data visualization, comprehensive early warning and reporting systems, and information management; The data processing layer (core functional module) is used to formulate early warning indicators, generate charts and conduct comprehensive early warning; The storage layer is used to manage historical data and real-time collection of data.

8. The earth-rock dam seepage safety monitoring, forecasting and early warning system according to claim 7, characterized in that: The security warning visualization module supports data flow and function linkage. After the user logs in, the real-time / historical seepage data and warning status can be viewed through the data visualization interface; The data processing layer calls data from the storage layer, completes the dynamic calculation of early warning indicators and generates charts, feeds the results back to the display layer, and realizes global risk integration through the comprehensive early warning module.

Images