A method for analyzing dam seepage stability based on flood forecasting

Abstract

The present invention discloses a method for analyzing dam seepage stability based on flood forecasting, which relates to the technical field of water conservancy project safety monitoring. The method includes S1, a data acquisition layer integrating a flood forecasting system and a seepage monitoring system, wherein sensors in the data acquisition layer acquire flood forecasting data and seepage monitoring data; S2, a spatiotemporal coupling interface of a dynamic fusion layer that aligns flood forecasting data with seepage monitoring data in spatiotemporal order, and generates dynamic driving variables for seepage field boundary conditions based on a flood evolution model and geographic information system spatial interpolation technology. The present invention achieves spatiotemporal alignment of flood forecasting data and seepage monitoring data and generation of dynamic driving variables through the spatiotemporal coupling interface, solving the problem in traditional analysis that boundary conditions are static or lagging and cannot reflect the real-time flood evolution process. The method enables the boundary conditions for seepage field calculation to be dynamically updated with flood forecasts, thereby improving the timeliness and accuracy of model calculations.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy project safety monitoring, and in particular to a dam seepage stability analysis method based on flood forecasting. Background Art

[0002] Seepage stability analysis of dams is a key technology for ensuring the safety of hydraulic projects. Its core lies in evaluating the evolution of key parameters such as seepage pressure and infiltration line through seepage field simulation and monitoring data inversion. Currently, the industry generally uses finite element numerical models based on Darcy's law. Seepage boundary conditions are set in conjunction with historical hydrological data, and stability is determined based on a standardized permeability coefficient inversion process. However, existing methods have significant limitations when addressing dynamic conditions such as floods. Firstly, traditional models rely on static design conditions and fail to effectively integrate real-time flood forecast data (such as flow rate and water level change rate), resulting in a disconnect between the seepage analysis boundary conditions and the actual flood evolution process. Secondly, seepage parameter correction often uses an offline iterative model, which has a long update cycle and is difficult to capture the dynamic changes in seepage characteristics during short-term heavy rainfall. Furthermore, information silos created by data protocol differences between flood forecasting systems and seepage monitoring systems prevent seepage risk assessments from providing a forward-looking view of the flood forecast period, hindering the timeliness of warnings and the reliability of decision-making.

[0003] Patent CN111382526B discloses a dam seepage analysis method that identifies anti-seepage section types and couples monitoring data. The above patent realizes real-time monitoring of the seepage pressure of earth-rock dams and timely grasps the safety status of earth-rock dams.

[0004] The above patent established a seepage calculation framework based on section classification by constructing an anti-seepage section type library, analytical method programmatic calculation and monitoring data coupling analysis. However, it relies on analytical calculation of preset anti-seepage section types and does not consider the real-time impact of dynamic boundary conditions such as water level and flow on the seepage field during the evolution of floods. It is difficult to reflect the seepage mutation characteristics under extreme flood conditions.

[0005] To this end, this application proposes a flood forecast-based dam seepage stability analysis method that can convert dynamic parameters such as flow and water level in the flood evolution process into real-time driving variables of seepage field boundary conditions by constructing a spatiotemporal coupling interface between flood forecast data and seepage monitoring data. Summary of the Invention

[0006] The purpose of the present invention is to provide a method for analyzing the seepage stability of dams based on flood forecasting, so as to solve the technical problems raised in the above-mentioned background technology. On the one hand, the traditional model relies on static design conditions and fails to effectively integrate real-time flood forecast data, resulting in the disconnection between the boundary conditions of the seepage analysis and the actual flood evolution process; on the other hand, the correction of seepage parameters mostly adopts an offline iterative mode, which has a long update cycle and is difficult to capture the dynamic changes of seepage characteristics in short-term heavy rainfall.

[0007] To achieve the above object, the present invention provides the following technical solution: a method for analyzing dam seepage stability based on flood forecasting, comprising the following steps:

[0008] S1. The flood forecasting system and the seepage monitoring system are integrated into a data acquisition layer. The sensors in the data acquisition layer acquire flood forecast data and seepage monitoring data. The flood forecast data includes flow series, water level series, and rainfall data within the forecast period. The seepage monitoring data includes real-time piezometer pressure data, piezometer infiltration line data, and displacement meter deformation data.

[0009] S2, the spatiotemporal coupling interface of the dynamic fusion layer aligns flood forecast data with seepage monitoring data in spatiotemporal order, and generates dynamic driving variables of seepage field boundary conditions based on the flood evolution model and geographic information system spatial interpolation technology;

[0010] S3, the model calculation layer embeds the improved seepage finite element model and imports the dynamic driving variable loading time-varying boundary conditions, and the parameter adaptive correction module connects the piezometer to update the permeability coefficient through the sliding window recursive algorithm;

[0011] S4. The application service layer generates a seepage stability risk level map based on the seepage field calculation results using the fuzzy comprehensive evaluation method.

[0012] Preferably, the data acquisition layer uniformly accesses multi-source heterogeneous data through a connection protocol analysis module, and the protocol analysis module has built-in drivers and interfaces that support cross-platform data interaction of hydrological data protocols, Internet of Things communication protocols and database interfaces.

[0013] Preferably, the construction of the spatiotemporal coupling interface includes: establishing a spatial mapping relationship between flood forecast data and seepage field nodes, using a time series alignment algorithm to match the time resolution of the flood forecast with the calculation step size of the seepage model, and generating a time-varying hydraulic head sequence and an infiltration rate sequence of the boundary nodes;

[0014] The spatial mapping relationship is based on the Thiessen polygon method to perform spatial interpolation on the discrete monitoring point data in the basin. The discrete monitoring point data are obtained by obtaining the spatial coordinates of the grid nodes in the seepage field.

[0015] Preferably, the improved seepage finite element model is a three-dimensional saturated-unsaturated seepage coupling model.

[0016] Preferably, the real-time correction process of the parameter adaptive correction module includes:

[0017] Construct a dynamic response surface model of permeability coefficient k, permeability pressure p, and permeability q;

[0018] Adopting a sliding window recursive optimization algorithm, the permeability coefficient is iteratively updated using 10-30 minutes of monitoring data, with an update frequency of no less than once per minute;

[0019] The corrected permeability coefficient is fed back to the finite element model in real time, triggering the incremental calculation of the model.

[0020] Preferably, the dynamic driving variables include the time-varying head value of the Dirichlet boundary condition and the time-varying flow value of the Neumann boundary condition;

[0021] The generation of time-varying water head based on Dirichlet boundary is done by analyzing the water level series H for flood forecasting. t Boundary node mapping is performed, and the outflow flow of the Neumann boundary is generated based on Darcy's law calculation.

[0022] Preferably, the evaluation indicators of the fuzzy comprehensive evaluation method include:

[0023] The ratio of the burial depth of the seepage line to the critical burial depth, the ratio of the seepage pressure gradient to the allowable seepage gradient of the soil, the proportion of the area where the seepage velocity exceeds the safety threshold, and the dam safety factor Fs calculated based on the effective stress principle.

[0024] Preferably, the analysis method further comprises outputting dynamic simulation results of the seepage field;

[0025] The simulation results include 3D visualization of the dynamic evolution curve of the infiltration line;

[0026] Seepage pressure cloud map and seepage vector distribution map;

[0027] Seepage pressure-time history curve of anti-seepage wall and weak interlayer.

[0028] Preferably, the flood forecasting system and seepage monitoring system include a four-layer architecture of data acquisition layer, dynamic fusion layer, model calculation layer and application service layer;

[0029] The data collection layer transmits data to the server through sensors. The server processes the data and uploads it to the dynamic fusion layer. The dynamic fusion layer pushes the data to the model calculation layer after fusion. The results of the model calculation layer are uploaded to the application service layer through files.

[0030] Preferably, the forecast period of the flood forecast data is 24-72 hours, the time resolution is 5-15 minutes, and the collection frequency of the seepage monitoring data is 1-5 minutes / time.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. This invention achieves spatiotemporal alignment of flood forecast data and seepage monitoring data and generates dynamic driving variables through a spatiotemporal coupling interface. This solves the problem of static or lagging boundary conditions in traditional analysis, which cannot reflect the real-time flood evolution process. It enables the boundary conditions of seepage field calculations to be dynamically updated with flood forecasts, improving the timeliness and accuracy of model calculations.

[0033] 2. The present invention uses a parameter adaptive correction module to achieve dynamic correction of seepage parameters and real-time simulation of the seepage field. This solves the problem of fixed permeability coefficients in traditional models that cannot adapt to changes in actual working conditions. It improves the adaptability of seepage field calculations to complex boundary conditions and changes in soil properties, and enhances the reliability of analysis results.

[0034] 3. This invention uses a fuzzy comprehensive evaluation method to achieve comprehensive evaluation of multiple indicators and visualize the dynamic process of the seepage field. This solves the problems of one-sided evaluation of a single indicator and non-intuitive results display, making it easier for decision makers to quickly understand the seepage status and risk distribution of the dam.

[0035] 4. The present invention realizes efficient integration and cross-platform interaction of multi-source data through the protocol analysis module, solves the problem of compatibility of data from different sources and in different formats in traditional methods, and improves system compatibility and data processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Schematic diagram of the overall structure of the analysis method of the present invention;

[0037] Figure 2 This is a schematic diagram of the dynamic data fusion process of the present invention;

[0038] Figure 3 This is a schematic diagram of the spatiotemporal coupling interface process of the present invention;

[0039] Figure 4 Schematic diagram of the sliding window recursive optimization algorithm of the present invention. DETAILED DESCRIPTION

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0041] See also Figure 1 、 Figure 2 、 Figure 3 and Figure 4 The present invention provides an embodiment: a method for analyzing dam seepage stability based on flood forecasting, comprising the following steps:

[0042] S1. The flood forecasting system and the seepage monitoring system are integrated into a data acquisition layer. The sensors in the data acquisition layer acquire flood forecast data and seepage monitoring data. The flood forecast data includes flow series, water level series, and rainfall data within the forecast period. The seepage monitoring data includes real-time piezometer pressure data, piezometer infiltration line data, and displacement meter deformation data.

[0043] S2, the spatiotemporal coupling interface of the dynamic fusion layer aligns flood forecast data with seepage monitoring data in spatiotemporal order, and generates dynamic driving variables of seepage field boundary conditions based on the flood evolution model and geographic information system spatial interpolation technology;

[0044] S3, the model calculation layer embeds the improved seepage finite element model and imports the dynamic driving variable loading time-varying boundary conditions, and the parameter adaptive correction module connects the piezometer to update the permeability coefficient through the sliding window recursive algorithm;

[0045] S4, the application service layer generates a seepage stability risk level map using a fuzzy comprehensive evaluation method based on the seepage field calculation results;

[0046] The analysis method further includes outputting dynamic simulation results of the seepage field; the simulation results include a three-dimensional visualized dynamic evolution curve of the seepage line; a seepage pressure cloud map and a seepage vector distribution map; and a seepage pressure-time history curve of the anti-seepage wall and the weak interlayer;

[0047] Furthermore, first, in the data collection stage: multiple types of sensors are deployed within the dam basin, including radar water level gauges and ultrasonic flow meters at the hydrological station to obtain the flow sequence Q(t), water level sequence H(t) and rainfall data transmitted by meteorological satellites within the forecast period of 24-72 hours; vibrating wire piezometers are buried every 50m in the dam body, a group of servo pressure measuring tubes are set every 30m along the dam axis, and static levels are set every 20m in key monitoring sections to collect real-time seepage pressure p, infiltration line buried depth h and dam body vertical displacement Δs. The data are transmitted to the edge server via a 4G wireless module at a frequency of 1 minute / time.

[0048] Then, the dynamic fusion layer starts the spatiotemporal coupling processing: based on the 10-minute water level and flow data output by the flood evolution model, the ArcGIS geographic information system is used to perform Thiessen polygon spatial interpolation on the seepage field grid nodes, and a weight mapping relationship is established between each node and the three surrounding monitoring points; through the time series alignment algorithm, the 10-minute resolution data of the flood forecast is resampled to the 5-minute step required by the seepage model, and the time-varying water head H of each boundary node is generated. n (t) and the infiltration rate vn (t) sequence, synchronously import the dynamic boundary condition database.

[0049] Finally, the model calculation layer loads the improved three-dimensional saturated-unsaturated seepage finite element model. The three-dimensional saturated-unsaturated seepage finite element model is developed using COMSOL Multiphysics. First, the initial permeability coefficient field assigned based on historical drilling data is imported, and then the H in the dynamic boundary condition database is read in real time. n (t) and v n (t), applying Dirichlet and Neumann boundaries; the parameter adaptive correction module is a parameter correction algorithm driven by time series. The parameter adaptive correction module calls the latest 30 minutes of piezometer data p1-p30 and pressure tube flow q1-q30 every 10 minutes, updates the permeability coefficient k(t) through a sliding window recursive algorithm, and the corrected k value automatically replaces the model parameters and triggers incremental calculation; after the calculation is completed, the service layer extracts results such as the infiltration line position and seepage pressure distribution, inputs them into the fuzzy comprehensive evaluation model, sets seven evaluation factors, and uses the hierarchical analysis method to determine the weights to generate a real-time risk level map. The risk level map is divided into four levels: green / yellow / orange / red, and is pushed to the flood control command center through the Web GIS platform.

[0050] See also Figure 1 、 Figure 2 and Figure 4 In one embodiment of the present invention, the flood forecasting system and seepage monitoring system includes a four-layer architecture: a data acquisition layer, a dynamic fusion layer, a model calculation layer, and an application service layer. The data acquisition layer transmits data to a server through sensors. The server processes the data and uploads it to the dynamic fusion layer. The dynamic fusion layer pushes the data to the model calculation layer after fusion. The results of the model calculation layer are uploaded to the application service layer via files.

[0051] The flood forecast data has a forecast period of 24-72 hours, a time resolution of 5-15 minutes, and a collection frequency of 1-5 minutes per seepage monitoring data;

[0052] Furthermore, the data acquisition layer supports hybrid networking using both wired and wireless communication. The dynamic fusion layer deploys a flood evolution model and a spatiotemporal interpolation weight allocation algorithm. The model calculation layer integrates a parallel computing engine, and the application service layer provides a visual interface with a B / S architecture, supporting real-time access from PCs and mobile devices. Regarding the hardware deployment of the data acquisition layer, a gateway supporting the SL651-2014 hydrological communication protocol is configured for the flood forecasting system's hydrological data, accessing XML-formatted flow and water level data from the National Hydrological Database. For the IoT devices in the seepage monitoring system, edge computing nodes supporting the MQTT protocol are deployed to parse binary data from piezometers connected via the Modbus-RTU protocol and pressure gauges connected via the RS485 interface in real time. Furthermore, a JDBC driver is used to connect to the reservoir operation database to obtain auxiliary data such as historical rainfall and evaporation.

[0053] Then, the protocol parsing module has a built-in multi-driver adaptation layer: for the hydrological data protocol, an XML parser is developed to extract fields such as timestamp, monitoring point number, and physical quantity value, and convert them into a unified JSON format, containing the "device ID-data type-time-value" key-value pair; for IoT devices, a protocol mapping table is designed. For example, the Modbus address 0x01 corresponds to the piezometer pressure value, and the conversion formula is P=0.01×register value. Real-time data conversion is achieved through the Node-RED flow platform; for the database interface, a timed polling mechanism is adopted to pull historical data every 5 minutes, and the data is cleaned by the ETL tool and stored in the time series database.

[0054] Finally, the unified access data is formatted on the edge server: the timestamp continuity and numerical rationality are checked, the timestamp allows an error of ±2 seconds, the water level value does not exceed ±20% of the historical extreme value, and abnormal data triggers a sensor fault warning; after the verification passes, the data is classified according to the "flood forecast / seepage monitoring / auxiliary data" label and pushed to the dynamic fusion layer through the Kafka message queue to ensure the consistency of multi-source heterogeneous data in timestamps and spatial coordinates, providing standardized input for subsequent spatiotemporal coupling processing.

[0055] See also Figure 1 、 Figure 2 and Figure 3 In one embodiment of the present invention, a method for analyzing dam seepage stability based on flood forecasting is provided. The data acquisition layer uniformly accesses multi-source heterogeneous data via a connection protocol parsing module. The protocol parsing module has built-in drivers and interfaces that support cross-platform data interaction using hydrological data protocols, Internet of Things communication protocols, and database interfaces.

[0056] The construction of the spatiotemporal coupling interface includes: establishing a spatial mapping relationship between flood forecast data and seepage field nodes, using a time series alignment algorithm to match the time resolution of the flood forecast with the calculation step size of the seepage model, and generating a time-varying head series and infiltration rate series of the boundary nodes; the spatial mapping relationship is based on the Thiessen polygon method to perform spatial interpolation on the discrete monitoring point data in the basin, and the discrete monitoring point data is obtained by obtaining the spatial coordinates of the grid nodes at the seepage field;

[0057] Furthermore, when constructing the spatial mapping relationship, high-precision DEM data of the dam and surrounding watershed are obtained through drone oblique photography. The three-dimensional grid is divided in the GMS seepage modeling software with a unit size of 10m×10m×5m, and the longitude and latitude coordinates (X, Y, Z) of each grid node are extracted. Based on the Thiessen polygon algorithm, the Thiessen polygons of the watershed monitoring points are generated in ArcGIS. Each grid node is automatically associated with the three nearest monitoring points, and the spatial interpolation weight is calculated, such as the inverse distance weighting method, the weight ω i =1 / d i / Σ1 / d i , the basin monitoring points include hydrological stations and seepage monitoring holes.

[0058] Then, the time series alignment is processed: the flow Q(t) and water level H(t) sequences output by the flood forecasting system are obtained, with a time resolution of 15 minutes, a forecast period of 72 hours, and a seepage model calculation step size of 5 minutes; the cubic spline interpolation algorithm is used to resample the 15-minute interval data into a continuous sequence of 5-minute intervals to generate Q'(t) and H'(t); at the same time, the seepage monitoring data is subjected to a sliding average filter through a filtering algorithm to reduce high-frequency noise, thereby obtaining the smoothed seepage pressure p(t) and flow q(t) sequences, and the acquisition frequency of the sliding average filter is 1 minute / time.

[0059] Finally, generate the boundary node driving variables: For the Dirichlet boundary nodes, calculate the time-varying water head H of each node according to H'(t) and the spatial interpolation weight n (t)= ×H i (t); For Neumann boundary nodes, based on Darcy's law Q = k × A × , combined with the real-time permeability coefficient k, calculate the time-varying outflow flow q n (t)=k×A n × , where An is the cross-sectional area of ​​the node, L n is the distance between upstream and downstream nodes, the initial value of the permeability coefficient k is the value in the survey report; n (t) and q n(t) The data is stored in the dynamic boundary condition table in the format of node number-time-value, which can be called by the model calculation layer in real time to achieve accurate spatiotemporal mapping of flood forecast data to seepage boundary conditions.

[0060] See also Figure 1 、 Figure 3 and Figure 4 , an embodiment provided by the present invention: a method for analyzing dam seepage stability based on flood forecasting, wherein the improved seepage finite element model is a three-dimensional saturated-unsaturated seepage coupling model;

[0061] The real-time correction process of the parameter adaptive correction module includes: constructing a dynamic response surface model of the permeability coefficient k, the permeability pressure p, and the permeability q; using a sliding window recursive optimization algorithm to iteratively update the permeability coefficient using 10-30 minutes of monitoring data, with an update frequency of not less than once per minute; and feeding the corrected permeability coefficient back to the finite element model in real time to trigger the incremental calculation of the model;

[0062] Furthermore, first, a dynamic response surface model was constructed: three groups of typical working conditions in the historical flood process, such as small / medium / large floods, were selected, the piezometer data p(t), the pressure tube flow q(t) and the true value of the manually inverted permeability coefficient k of the corresponding period were collected, and multiple regression analysis was used to establish k=f(p, q, d t ) empirical formula, initially assuming k=α p +β q +γ, where α, β, and γ are regression coefficients. For complex nonlinear relationships, an LSTM neural network model is introduced, which inputs the p and q sequences of the previous 30 minutes and outputs the current k value. The model is trained using historical data until the error is ≤5%.

[0063] Then, the real-time correction process is carried out: the parameter adaptive correction module uses a 10-minute sliding window, extracts the latest set of monitoring data (p(t), q(t)) every 1 minute, adds the window and removes the earliest set of data, keeping the window always filled with the 10 sets of data from the last 10 minutes; uses the recursive least squares method to iteratively optimize the regression model parameters α, β, and γ, or performs online fine-tuning on the LSTM model. The online fine-tuning is triggered every 10 minutes, collects the latest 30 minutes of monitoring data, forms a sliding window sequence, and calculates the current permeability coefficient k(t); sets the update frequency to once per minute, that is, updates the k value at least 5 times in each calculation step to ensure that the sudden change in permeability characteristics caused by short-term heavy rainfall is captured in a timely manner.

[0064] Finally, a feedback mechanism for the revised results was developed: when the difference between k(t) and the previous k(t-1) exceeds 10%, an incremental calculation of the finite element model is triggered. This involves maintaining the meshing unchanged, updating only the permeability coefficient in the material properties, and continuing the solution from the current calculation step to avoid time-consuming recalculation of the entire model. At the same time, k(t) is stored in the parameter history library for subsequent trend analysis. If the k value increases sharply three times in a row, it is automatically marked as a "piping risk warning." The parameter adaptive correction module generates a warning event data packet and encapsulates it in JSON format via the RESTful API interface. A digital signature is attached to ensure data integrity, and the data is pushed to the application service layer after encryption via the HTTPS protocol. The application service layer triggers a red flashing alarm on the Web GIS visualization interface, locates the abnormal area, and highlights the corresponding seepage vector distribution map. The left side of the manual review interface displays the real-time seepage pressure curve and k value fluctuation trend, the right side loads historical CT scan images of the area, and a link to the expert knowledge base is provided below. Staff can use the manual review interface to view the raw sensor data from the past hour.

[0065] See also Figure 1 、 Figure 3 and Figure 4 The present invention provides an embodiment: a method for analyzing dam seepage stability based on flood forecasting, wherein the dynamic driving variables include the time-varying head value of the Dirichlet boundary condition and the time-varying flow value of the Neumann boundary condition; the generation of the time-varying head of the Dirichlet boundary is achieved by analyzing the water level sequence H of the flood forecasting. t Perform boundary node mapping, and generate the outflow flow of Neumann boundary based on Darcy's law calculation;

[0066] The evaluation indicators of the fuzzy comprehensive evaluation method include: the ratio of the depth of the infiltration line to the critical depth, the ratio of the seepage pressure gradient to the allowable seepage gradient of the soil, the proportion of the area where the seepage velocity exceeds the safety threshold, and the dam safety factor F calculated based on the effective stress principle. s ;

[0067] Furthermore, first, the dynamic evolution of the seepage line is visualized in three dimensions: after each seepage field solution is completed in the model calculation layer, the buried depth h(t) of the seepage line at each piezometer position is extracted. Combined with the GIS geographic coordinates, a three-dimensional model of the dam is generated on the Unity3D platform, and the position of the seepage line is rendered in real time with a red curve. A timeline control is used to support users to drag and view the dynamic changes of the seepage line in the past 24 hours. When h(t) approaches the critical buried depth, the curve flashes yellow as a warning. The critical buried depth is calculated according to the soil mechanics formula, such as h 临界 =H 上游 × .

[0068] Then, the seepage pressure cloud map and vector distribution are drawn: using the Para View open source software, the node pressure values ​​obtained by finite element calculation are interpolated into a continuous cloud map, the upstream area adopts a blue to red gradient (blue <0.1MPa, red >0.5MPa), and the seepage vector arrow is superimposed near the downstream escape point. The size of the arrow reflects the flow rate, and the direction indicates the seepage path; when the seepage velocity in a certain area exceeds the safety threshold, the red grid unit is automatically marked, and the proportion of the area exceeding the standard is counted as one of the indicators of fuzzy evaluation.

[0069] Finally, extract the specific structure seepage pressure-time history: for the anti-seepage wall, such as the concrete anti-seepage wall, arranged along the dam axis, a monitoring point is preset every 10m in the model, and the seepage pressure P at each point is extracted. wall (t), generate a broken line graph and superimpose the historical maximum value and warning threshold; for weak interlayers, such as clay interlayers, buried at a depth of 5-8m, extract the pressure difference Δp(t) between the top and bottom surfaces of the interlayer and calculate the seepage gradient I= , when I exceeds the allowable slope I 允 When the interlayer penetration damage warning is triggered, I 允 =0.8×I 临界 The simulation results are provided in the form of a Web API, supporting real-time loading on mobile devices, and PDF report archives are generated, which include 1-hour / 6-hour / 24-hour trend analysis.

[0070] Working Principle: First, the data acquisition layer uses sensors to acquire multi-source heterogeneous data, including flow series, water level series, and rainfall data within the forecast period of the flood forecasting system, as well as real-time pressure data from the piezometer, piezometer infiltration line data, and displacement meter deformation data from the seepage monitoring system. These data are then uniformly accessed through a protocol parsing module to achieve cross-platform data interaction. The spatiotemporal coupling interface of the dynamic fusion layer aligns these data in time and space. Based on the flood evolution model and geographic information system spatial interpolation technology, a spatial mapping relationship between flood forecast data and seepage field nodes is established. A time series alignment algorithm is used to match the flood forecast time resolution with the seepage model calculation step size, generating dynamic driving variables for the seepage field boundary conditions, such as the time-varying head value at the upstream boundary and the time-varying flow value at the downstream boundary, providing real-time dynamic boundary conditions for seepage field analysis.

[0071] Then, seepage model calculations and real-time parameter corrections are carried out. The model calculation layer embeds an improved seepage finite element model for three-dimensional saturated-unsaturated seepage coupling, and imports dynamic drive variable loading time-varying boundary conditions. At the same time, the parameter adaptive correction module constructs a dynamic response surface model of the permeability coefficient k, seepage pressure p, and seepage volume q. Using 10-30 minutes of monitoring data and a sliding window recursive optimization algorithm, the permeability coefficient is iteratively updated at a frequency of no less than once per minute. The corrected coefficient is fed back to the finite element model in real time, triggering the model's incremental calculations. This achieves adaptive optimization of the permeability coefficient during flooding and improves the ability to capture changes in permeability characteristics under short-term heavy rainfall conditions.

[0072] Finally, the application service layer uses the dynamic simulation results of the infiltration line, seepage pressure, etc. obtained based on the seepage field calculation, including the three-dimensional visualization of the dynamic evolution curve of the infiltration line, the seepage pressure cloud map and the seepage vector distribution map, the seepage pressure-time history curve of the anti-seepage wall and the weak interlayer, etc., and adopts the fuzzy comprehensive evaluation method, combined with the ratio of the infiltration line burial depth to the critical burial depth, the ratio of the seepage pressure gradient to the allowable seepage gradient of the soil, the proportion of the area where the seepage velocity exceeds the safety threshold, the dam safety factor Fs calculated based on the effective stress principle and other evaluation indicators to generate the seepage stability risk level map and early warning instructions, providing real-time and accurate decision-making support for flood control and rescue.

[0073] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for analyzing dam seepage stability based on flood forecasting, characterized by: The following steps are involved: S1. The flood forecasting system and the seepage monitoring system are integrated into a data acquisition layer. The sensors in the data acquisition layer acquire flood forecast data and seepage monitoring data. The flood forecast data includes flow series, water level series, and rainfall data within the forecast period. The seepage monitoring data includes real-time piezometer pressure data, piezometer infiltration line data, and displacement meter deformation data. The spatiotemporal coupling interface of the S2 dynamic fusion layer aligns flood forecast data with seepage monitoring data in spatiotemporal order, and generates dynamic driving variables for seepage field boundary conditions based on the flood evolution model and geographic information system spatial interpolation technology. The cubic spline interpolation algorithm is used to resample the 15-minute interval data into a continuous sequence of 5-minute intervals to generate Q'(t) and H'(t). For Dirichlet boundary nodes, the time-varying hydraulic head of each node is calculated based on H'(t) and spatial interpolation weights. For Neumann boundary nodes, based on Darcy's law Combined with the real-time permeability coefficient k, calculate the time-varying outflow flow Where An is the cross-sectional area of ​​the node, L n is the distance between upstream and downstream nodes, the initial value of the permeability coefficient k is the value in the survey report; n (t) and q n (t) Store the dynamic boundary condition table in the node number-time-value format for real-time call by the model calculation layer; S3, the model calculation layer embeds the improved seepage finite element model and imports the dynamic driving variable loading time-varying boundary conditions, and the parameter adaptive correction module connects the piezometer to update the permeability coefficient through the sliding window recursive algorithm; S4. The application service layer generates a seepage stability risk level map based on the seepage field calculation results using the fuzzy comprehensive evaluation method.

2. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The data acquisition layer uniformly accesses multi-source heterogeneous data through a connection protocol analysis module. The protocol analysis module has built-in drivers and interfaces that support cross-platform data interaction of hydrological data protocols, Internet of Things communication protocols, and database interfaces.

3. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The construction of the spatiotemporal coupling interface includes: establishing a spatial mapping relationship between flood forecast data and seepage field nodes, using a time series alignment algorithm to match the time resolution of the flood forecast with the calculation step size of the seepage model, and generating a time-varying hydraulic head series and an infiltration rate series of the boundary nodes; The spatial mapping relationship is based on the Thiessen polygon method to perform spatial interpolation on the discrete monitoring point data in the basin. The discrete monitoring point data are obtained by obtaining the spatial coordinates of the grid nodes in the seepage field.

4. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The improved seepage finite element model is a three-dimensional saturated-unsaturated seepage coupling model.

5. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The real-time correction process of the parameter adaptive correction module includes: Construct a dynamic response surface model of permeability coefficient k, permeability pressure p, and permeability q; Adopting a sliding window recursive optimization algorithm, the permeability coefficient is iteratively updated using 10-30 minutes of monitoring data, with an update frequency of no less than once per minute; The corrected permeability coefficient is fed back to the finite element model in real time, triggering the incremental calculation of the model.

6. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The dynamic driving variables include the time-varying head value of the Dirichlet boundary condition and the time-varying flow value of the Neumann boundary condition; The generation of the time-varying hydraulic head of the Dirichlet boundary is achieved by mapping the boundary nodes of the water level series H(t) for flood forecasting, and the generation of the outflow flow of the Neumann boundary is based on the calculation of Darcy's law.

7. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The evaluation indicators of the fuzzy comprehensive evaluation method include: The ratio of the burial depth of the seepage line to the critical burial depth, the ratio of the seepage pressure gradient to the allowable seepage gradient of the soil, the proportion of the area where the seepage velocity exceeds the safety threshold, and the dam safety factor F calculated based on the effective stress principle. s .

8. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The analysis method further includes outputting dynamic simulation results of the seepage field; The simulation results include 3D visualization of the dynamic evolution curve of the infiltration line; Seepage pressure cloud map and seepage vector distribution map; Seepage pressure-time history curve of anti-seepage wall and weak interlayer.

9. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The flood forecasting system and seepage monitoring system include a four-layer architecture of data acquisition layer, dynamic fusion layer, model calculation layer and application service layer; The data collection layer transmits data to the server through sensors. The server processes the data and uploads it to the dynamic fusion layer. The dynamic fusion layer pushes the data to the model calculation layer after fusion. The results of the model calculation layer are uploaded to the application service layer through files.

10. The method for analyzing dam seepage stability based on flood forecasting according to claim 1, characterized in that: The flood forecast data has a forecast period of 24-72 hours, a time resolution of 5-15 minutes, and a collection frequency of seepage monitoring data of 1-5 minutes per time.

Images