Seepage parameter inversion prediction method and system for core wall rock-fill dam

Abstract

This invention relates to the field of dam seepage technology, and particularly to a method and system for inverting and predicting seepage parameters in a core-rockfill dam. The method includes: acquiring a target seepage performance score for the dam body; collecting operational data of the core-rockfill dam at preset acquisition time intervals to obtain a set of dam body seepage characteristic data; analyzing the acquired multiple sets of dam body seepage characteristic data to obtain a real-time seepage performance score for the dam body; determining whether seepage inversion prediction is necessary based on the real-time seepage performance score and the target seepage performance score; and, if seepage inversion prediction is necessary, determining the model parameters of a preset seepage parameter inversion prediction model; and inputting the latest acquired set of dam body seepage characteristic data into the seepage parameter inversion prediction model with confirmed model parameters to obtain the inversion prediction result of the core-rockfill dam seepage parameters. This improves the accuracy of seepage parameter inversion prediction.

Description

Technical Field

[0001] This invention relates to the technical field of dam seepage, and in particular to a method and system for inverting and predicting seepage parameters of a core rockfill dam. Background Technology

[0002] Core-wall rockfill dams are a widely used dam type in water conservancy projects. Their seepage safety is directly related to the stability of the dam structure and its long-term service performance. Seepage parameters are the core indicators for evaluating the seepage state of the dam body. Changes in seepage parameters may cause problems such as leakage and seepage deformation of the dam body, and in severe cases, they may even threaten the safety of the dam. The inverse prediction of seepage parameters aims to reverse the seepage characteristics inside the dam body by monitoring data, and provide a scientific basis for the safety of the project.

[0003] Existing inversion methods typically rely on historical data or calibration parameters under specific operating conditions, lacking the ability to dynamically respond to real-time operating data. When the dam body encounters extreme operating conditions or local damage, existing models struggle to adjust parameters in a timely manner, leading to delayed or deviated prediction results from actual needs. Summary of the Invention

[0004] This invention provides a method for predicting seepage parameters in core-wall rockfill dams, which can improve the accuracy of seepage parameter inversion prediction and effectively solve the problems in the background art.

[0005] To achieve the above objectives, this invention provides a method for inverting and predicting seepage parameters in core-wall rockfill dams, including:

[0006] Obtain the target seepage performance score of the dam body;

[0007] According to the preset collection time interval, the operation data of the core rockfill dam is collected to obtain a set of seepage characteristic data of the dam body;

[0008] Based on the acquired multiple sets of dam seepage characteristic data, a real-time seepage performance score of the dam body is obtained through analysis.

[0009] Based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body, it is determined whether seepage inversion prediction is needed; and if seepage inversion prediction is needed, the model parameters of the preset seepage parameter inversion prediction model are determined.

[0010] The newly acquired dam seepage characteristic data set is input into the seepage parameter inversion prediction model with confirmed model parameters to obtain the inversion prediction results of the seepage parameters of the core rockfill dam.

[0011] Furthermore, obtain the target seepage performance score of the dam body, including:

[0012] Obtain the design standards for the core-wall rockfill dam to be inverted, identify and extract its seepage parameter data, and obtain the target seepage parameter set;

[0013] The target seepage parameter set is analyzed for seepage performance to obtain the target seepage performance score of the dam body.

[0014] Furthermore, the target seepage parameter set includes the target permeability coefficient and the target hydraulic gradient.

[0015] Furthermore, the dam seepage characteristic data set includes the water head, seepage flow rate, and pore water pressure at a set location on the dam body.

[0016] Furthermore, based on the acquired multiple sets of dam seepage characteristic data, a real-time seepage performance score for the dam body is obtained through analysis, including:

[0017] According to the order of the collection timestamps, a set number of dam seepage characteristic data sets are sorted and aligned to obtain the dam seepage characteristic matrix;

[0018] The seepage environment characteristic matrix of the dam body is input into the pre-constructed seepage performance evaluation model to obtain the real-time seepage performance score of the dam body.

[0019] Furthermore, based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body, it is determined whether seepage inversion prediction is needed; and if seepage inversion prediction is needed, the model parameters of the preset seepage parameter inversion prediction model are determined, including:

[0020] The deviation of the seepage performance is calculated based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body.

[0021] Compare the deviation in seepage performance with the preset error range:

[0022] If the deviation in seepage performance is within the preset error range, it indicates that the current seepage state is close to the ideal state, and there is no need to perform inversion prediction.

[0023] If the deviation in seepage performance exceeds the preset error range, seepage inversion prediction is required, and the model parameters of the preset seepage parameter inversion prediction model are determined based on the deviation in seepage performance.

[0024] Furthermore, the seepage performance evaluation model is constructed using a multilayer perceptron architecture, where the number of nodes in the input layer is equal to the dimension of the seepage feature matrix, the hidden layer contains neurons with ReLU activation functions, and the output layer generates standardized score values ​​through the Sigmoid function.

[0025] Furthermore, the preset data collection time interval is dynamically adjusted based on the environmental conditions of the dam, seasonal changes, and extreme working conditions monitored in real time.

[0026] Furthermore, the preset error range is determined comprehensively based on the historical operating data, design standards, and safety margin of the core rockfill dam, and is periodically optimized and adjusted based on the latest acquired operating data.

[0027] On the other hand, the present invention also provides a seepage parameter inversion and prediction system for core-wall rockfill dams, comprising:

[0028] The target score acquisition module is used to acquire the target seepage performance score of the dam body;

[0029] The data acquisition module is used to collect operational data of the core rockfill dam according to a preset acquisition time interval, and obtain a set of seepage characteristic data of the dam body;

[0030] The real-time analysis module is used to analyze and obtain a real-time seepage performance score of the dam body based on multiple sets of acquired dam body seepage characteristic data;

[0031] The decision module is used to determine whether seepage inversion prediction is needed based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body; and when seepage inversion prediction is needed, to determine the model parameters of the preset seepage parameter inversion prediction model.

[0032] The inversion prediction module is used to input the latest acquired dam seepage characteristic data set into the seepage parameter inversion prediction model with confirmed model parameters to obtain the inversion prediction results of seepage parameters for core rockfill dams.

[0033] The technical solution of this invention achieves the following technical effects: Real-time acquisition of dam operation data is achieved by collecting dam seepage characteristic data at preset acquisition time intervals; real-time seepage performance scores are obtained based on dam operation data, which, compared to traditional methods relying on historical data or specific operating conditions for parameter calibration, can quickly respond to changes in the dam's seepage state, especially under extreme operating conditions or local damage, enabling timely detection of seepage anomalies and avoiding lag in prediction results; comparing the real-time seepage performance score with the target seepage performance score determines whether to perform seepage inversion prediction, avoiding unnecessary waste of computational resources and improving efficiency; inversion prediction is only initiated when the seepage performance significantly deviates from the target, ensuring that the inversion prediction addresses actual needs and concentrates resources on key issues; the parameters of the seepage parameter inversion prediction model are determined based on the judgment results, allowing the model to better adapt to the current actual situation of the dam; inputting the latest data into the model with confirmed parameters fully utilizes real-time information, improving the accuracy of seepage parameter inversion prediction and providing a more reliable scientific basis for dam safety assessment and decision-making. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart of the seepage parameter inversion and prediction method for the central wall rockfill dam of the present invention;

[0036] Figure 2 This is a structural block diagram of the seepage parameter inversion and prediction system for the central wall rockfill dam of the present invention. Detailed Implementation

[0037] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0038] like Figure 1 As shown, the method for inverting and predicting seepage parameters of a core-wall rockfill dam according to the present invention specifically includes the following steps:

[0039] Step S1: Obtain the target seepage performance score of the dam body;

[0040] Step S2: Collect operational data of the core-wall rockfill dam according to the preset collection time interval to obtain a set of dam seepage characteristic data; the set of dam seepage characteristic data includes water head, seepage flow rate and pore water pressure at a set location on the dam body;

[0041] Step S3: Based on the acquired multiple sets of dam seepage characteristic data, analyze and obtain a real-time seepage performance score for the dam body;

[0042] Step S4: Based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body, determine whether seepage inversion prediction is needed; and if seepage inversion prediction is needed, determine the model parameters of the preset seepage parameter inversion prediction model.

[0043] Step S5: Input the newly acquired dam seepage characteristic data set into the seepage parameter inversion prediction model with confirmed model parameters to obtain the inversion prediction results of the seepage parameters of the core rockfill dam.

[0044] In this embodiment, the technical solution of the present invention can achieve the following technical effects: By collecting dam seepage characteristic data at preset acquisition time intervals, real-time dam operation data can be obtained; based on the dam operation data, a real-time seepage performance score is obtained. Compared with traditional methods that rely on historical data or specific operating conditions to calibrate parameters, this method can quickly respond to changes in the dam's seepage state, especially under extreme operating conditions or local damage, enabling timely detection of seepage anomalies and avoiding lag in prediction results; by comparing the real-time seepage performance score of the dam with the target seepage performance score, it is determined whether to perform seepage inversion prediction, avoiding unnecessary waste of computational resources and improving efficiency; inversion prediction is only initiated when the seepage performance deviates significantly from the target, ensuring that the inversion prediction is tailored to actual needs and that resources are concentrated on key issues; the parameters of the seepage parameter inversion prediction model are determined based on the judgment results, allowing the model to better adapt to the current actual situation of the dam; inputting the latest data into the model with confirmed parameters fully utilizes real-time information, improves the accuracy of seepage parameter inversion prediction, and provides a more reliable scientific basis for dam safety assessment and decision-making.

[0045] Furthermore, the above-mentioned method achieves a systematic upgrade of traditional seepage analysis methods through a closed-loop architecture of "dynamic monitoring-intelligent evaluation-adaptive inversion". The synergistic effect of each link produces a significant composite effect. By continuously collecting operational data and comparing real-time and target seepage performance scores, it can keenly capture seepage anomalies in the dam body. The whole system forms a closed loop from data collection, anomaly judgment, model adaptation to accurate prediction, effectively overcoming the drawbacks of traditional methods that rely on historical data and have a delayed response.

[0046] In some embodiments of the present invention, the method for obtaining the target seepage performance score of the dam body specifically includes:

[0047] Step S11: Obtain the design standards for the core-wall rockfill dam to be inverted. Each core-wall rockfill dam will have a complete set of design standards developed during the design phase, based on relevant engineering specifications, geological conditions, hydrological conditions, and expected functional requirements. The design standards include parameter requirements for dam design and operation. Specifically, the design standards include the following:

[0048] Dam dimensions and structural type: such as the dam's height, width, length and other geometric parameters, as well as the specific rockfill materials and technologies used;

[0049] Hydraulic parameters: such as design head, maximum allowable seepage flow, pore water pressure distribution, etc.

[0050] Safety indicators: such as the safety factor of the design, the maximum allowable deformation, etc.

[0051] Step S12: Identify and extract seepage parameter data from the design standards to obtain the target seepage parameter set. The target seepage parameter set consists of the seepage-related parameters identified and extracted from the design standards. The target seepage parameter set mainly includes the target permeability coefficient and the target hydraulic gradient. The permeability coefficient reflects the ability of the dam material to allow water to pass through, usually in cm / s, and is an important parameter for calculating the seepage path and velocity. The hydraulic gradient represents the pressure gradient when water flows through the dam body and is used to evaluate the hydrodynamic characteristics at different locations inside the dam body.

[0052] Step S13: Analyze the seepage performance of the target seepage parameter set to obtain the target seepage performance score of the dam body. After obtaining the target seepage parameter set, a comprehensive evaluation is conducted using seepage theory and analysis methods. This includes establishing a mathematical model based on the principles of seepage mechanics, simulating the seepage situation of the dam body under different working conditions using computer simulation software, or combining practical experience and theoretical knowledge from similar projects in the past to conduct qualitative and quantitative analysis and judgment of the target seepage parameters. Based on the results of the seepage performance analysis, the target seepage performance score of the dam body is determined according to pre-set scoring rules or standards. The scoring rules are derived from a large number of experimental studies, engineering case statistics, and expert experience summaries, comprehensively considering the degree of influence of seepage parameters on the safety and stability of the dam body. For example, the closer the target permeability coefficient and target hydraulic gradient are to the ideal value or within the safe range, the better the seepage performance of the dam body, and the higher the corresponding score. Conversely, if these parameters deviate significantly from the ideal value and exceed the safe range, it indicates that there may be potential seepage risks in the dam body, and the score will be lower.

[0053] In this embodiment, starting with the design standards of the rockfill dam with core wall to be inverted, the process of obtaining the target seepage performance score can comprehensively consider various key elements of dam design and operation, avoiding the problem of incomplete evaluation due to focusing on only a single factor, thus laying a solid foundation for the subsequent accurate acquisition of seepage performance scores. By specifically identifying and extracting seepage parameter data from the design standards, the focus is clearly on the two key parameters directly related to seepage: the target permeability coefficient and the target hydraulic gradient, forming a target seepage parameter set. This eliminates interference from irrelevant information, making the subsequent analysis process more targeted and improving the efficiency and accuracy of obtaining the target seepage performance score of the dam.

[0054] In some embodiments of the present invention, the dam seepage characteristic data set includes the water head, seepage flow rate, and pore water pressure at a set location on the dam body;

[0055] Water head refers to the energy possessed by a unit weight of liquid. In the context of dam seepage, the water head at a designated location on the dam represents the energy height of the water at that specific location, measured relative to a reference plane, such as the bottom of the dam foundation. Water head is the driving force behind seepage. Differences in water head at different locations cause water to flow within the dam, and its magnitude and distribution directly affect parameters such as seepage velocity, direction, and flow rate. By monitoring the water head at a designated location on the dam, we can understand the distribution of water head within the dam, analyze the driving force of seepage, and determine whether there are seepage problems that may be caused by unreasonable water head distribution, such as excessively high local seepage velocities.

[0056] Seepage flow refers to the amount of water passing through a specific cross-section or area of ​​a dam per unit time, usually expressed in cubic meters per second (m³ / s). 3 The volumetric flow rate is expressed in units such as / s; seepage flow rate is an important indicator for measuring the degree of seepage in a dam; under normal circumstances, the seepage flow rate of a dam should be controlled within a reasonable range; if the seepage flow rate suddenly increases or continues to exceed the design allowable value, it may mean that there are seepage channels in the dam, changes in the properties of the dam materials, or damage to the dam structure; therefore, real-time monitoring of seepage flow rate is crucial for timely detection of potential seepage hazards in the dam and for assessing the safety of the dam.

[0057] Pore ​​water pressure refers to the pressure exerted by water within the pores of a dam. In granular dam structures such as earth-rock dams, pore water pressure significantly impacts the dam's stability and seepage characteristics. It is typically expressed as a pressure value relative to atmospheric pressure. Changes in pore water pressure affect the effective stress between dam particles, thereby influencing the dam's shear strength and stability. Increased pore water pressure reduces the effective stress between dam particles, lowering the dam's shear strength and increasing the risk of instability such as sliding and collapse. When pore water pressure data differs from the expected value obtained through simulation analysis based on the target seepage parameter set, it can be used to determine whether the dam is in a stable seepage state. This provides crucial information for evaluating the dam's real-time seepage performance, helping to promptly identify potential seepage safety hazards and take appropriate measures to ensure the safe operation of the dam.

[0058] In this embodiment, by clearly defining and defining the roles of three core parameters—head, seepage flow rate, and pore water pressure—in the dam seepage characteristic data set, a multi-dimensional and high-precision seepage monitoring system is constructed. Its advantages are as follows: First, head, as a core parameter of seepage dynamics, directly reflects the distribution of seepage driving force by quantifying energy differences at different locations. Combined with spatial monitoring, it can accurately identify local head anomalies, providing dynamic data support for analyzing seepage paths and velocities. Second, seepage flow rate, as an intuitive quantitative indicator, can quickly capture abnormal seepage exceeding design thresholds by monitoring changes in seepage scale in real time. Combined with historical data comparison, it can determine material aging. The monitoring of structural damage levels significantly improves the timeliness of hazard identification. Furthermore, pore water pressure monitoring combines seepage analysis with dam stability, assessing the risk of shear strength attenuation through effective stress changes. This overcomes the shortcomings of traditional methods that focus only on flow rate while neglecting mechanical effects. Simultaneously, by dynamically comparing the simulated values ​​with the target parameters, it achieves multi-criteria joint diagnosis of seepage status. In addition, the coordinated acquisition and fusion of the three types of parameters in the spatiotemporal dimensions form a closed-loop analysis chain from "driving force - seepage scale - structural response," ensuring that the data covers the key links of the entire seepage process and provides high-information-density input for model inversion, balancing monitoring comprehensiveness and computational efficiency.

[0059] In some embodiments of the present invention, a method for obtaining a real-time seepage performance score of a dam body includes:

[0060] After acquiring multiple sets of dam seepage characteristic data (including head, seepage flow, and pore water pressure), the first step is to process this data. Since the data changes dynamically over time, it needs to be sorted according to the timestamps of the data collection. This step ensures the temporal continuity and logical consistency of the data. To guarantee the effectiveness of the analysis, a reasonable amount of data is usually selected for analysis; for example, data from the past 24 hours or one week can be chosen to capture sufficient trends. Different types of seepage characteristic data may be collected at different times, so these data need to be aligned. The aligned data can form a unified time series, facilitating subsequent analysis. A structured dam seepage characteristic matrix can be obtained, containing information such as head, seepage flow, and pore water pressure at various locations within the dam over a specific time period.

[0061] The processed dam seepage feature matrix is ​​input into a pre-built seepage performance evaluation model to obtain a real-time seepage performance score for the dam. The seepage performance evaluation model is a pre-trained mathematical model used to evaluate the current seepage state of the dam based on the input seepage feature data. The seepage performance evaluation model can be built based on various technologies, such as machine learning algorithms, numerical simulation methods, or other statistical analysis methods. The seepage feature matrix, as input data, is calculated and analyzed within the model to finally output a standardized score value, namely the real-time seepage performance score of the dam. The real-time seepage performance score of the dam reflects the degree of deviation of the current seepage state of the dam from the ideal state.

[0062] On the other hand, the seepage performance evaluation model can also be constructed using a multilayer perceptron architecture, where the number of nodes in the input layer is equal to the dimension of the seepage feature matrix, the hidden layer contains neurons with ReLU activation functions, and the output layer generates standardized score values ​​through the Sigmoid function.

[0063] In this embodiment, data consistency and accuracy are ensured by sorting and aligning the data over time, avoiding errors caused by time inconsistencies. Selecting a reasonable time period for analysis can capture sufficient trends and improve the reliability and stability of the evaluation results. The model, built based on seepage theory and a large amount of experimental data, has high scientific validity and accuracy. The model can quantitatively evaluate complex seepage phenomena and provide objective scoring results. The model not only considers the changes of a single parameter but also combines data from multiple dimensions such as head, seepage flow, and pore water pressure to comprehensively reflect the seepage state of the dam.

[0064] In some embodiments of the present invention, step S4 triggers inversion prediction through dynamic deviation analysis and threshold decision mechanism, solving the problem that model updates in traditional methods rely on fixed cycles or manual intervention; by calculating the deviation between real-time scoring and target scoring and comparing the error range, adaptive triggering of inversion prediction is achieved, and model parameters are optimized based on the degree of deviation, thereby improving the response speed and prediction accuracy under extreme conditions. The specific implementation is as follows:

[0065] The deviation in seepage performance is calculated based on the real-time seepage performance score and the target seepage performance score of the dam body. The target seepage performance score is derived from the design standards of core-wall rockfill dams through rigorous scientific methods, such as using seepage theory, mathematical models, and simulation software, or combining past engineering experience to analyze the target seepage parameter set, reflecting the seepage performance of the dam body under ideal design conditions. The real-time seepage performance score of the dam body is obtained by collecting seepage characteristic data (such as water head, seepage flow, and pore water pressure at a set location on the dam body) during real-time operation. The data is processed and analyzed using a pre-built seepage performance evaluation model, which reflects the actual seepage performance of the dam body at present. By calculating the difference between the two, the seepage performance deviation is obtained. The seepage performance deviation can intuitively reflect the degree of deviation between the current seepage state of the dam body and the design expectation. For example, if the target seepage performance score of the dam body is 90 points (assuming that the full score is 100 points to represent the ideal state), and the real-time seepage performance score of the dam body is 70 points, then the seepage performance deviation is 20 points, indicating that there is a certain gap between the current seepage state and the ideal state.

[0066] After obtaining the deviation in seepage performance, it needs to be compared with the preset error range. The preset error range is determined comprehensively based on the historical operating data, design standards, and safety margins of the core-wall rockfill dam, and is periodically optimized and adjusted based on newly acquired operating data. The preset error range provides a reasonable reference interval for judging the seepage state of the dam. If the deviation in seepage performance is within the preset error range, it means that although the current seepage state of the dam differs from the target state to some extent, it is still within an acceptable range and close to the ideal state. In this case, there is no need to perform seepage inversion prediction. This is because the seepage situation of the dam is relatively stable under these circumstances, and there is no obvious indication that it is necessary to perform inversion prediction of seepage parameters to adjust the model. However, if the deviation in seepage performance exceeds the preset error range, it means that the current seepage state of the dam deviates significantly from the design expectation, and there may be potential seepage problems, such as the risk of leakage or seepage deformation of the dam. In this case, seepage inversion prediction is required.

[0067] After determining the need for inversion prediction, the model parameters of the preset seepage parameter inversion prediction model are determined based on the deviation of seepage performance. When determining the model parameters, the physical mechanisms of the seepage process are fully considered. For example, based on basic seepage principles such as Darcy's law (the flow rate is proportional to the hydraulic gradient when water seeps in a porous medium), the relationship between the deviation and factors such as the water velocity and pressure distribution inside the dam body is analyzed. If the deviation is large and shows an upward trend, it means that the water velocity or pressure distribution inside the dam body has changed significantly, and the parameters representing these physical quantities in the model need to be adjusted accordingly. Mathematical optimization algorithms are used to determine the model parameters; common methods include least squares, genetic algorithms, and particle swarm optimization. Taking least squares as an example, its basic idea is to adjust the model parameters to minimize the sum of squared errors between the model prediction results and the actual observed data. The specific process is as follows:

[0068] Using the deviation in seepage performance and related observation data (such as head, seepage flow, pore water pressure, etc. in the dam seepage characteristic data set) as input, an objective function is constructed, which represents the degree of difference between the model prediction and the actual observation.

[0069] Set a set of initial values ​​for the model parameters. These initial values ​​can be obtained based on experience, historical data, or simple theoretical calculations.

[0070] By continuously adjusting the model parameters and recalculating the value of the objective function, the objective function reaches its minimum value or meets certain convergence conditions. In each iteration, the direction and step size of parameter adjustment are determined based on the gradient information of the objective function (reflecting the direction of the fastest change in the objective function), gradually approaching the optimal solution.

[0071] In this embodiment, by calculating the seepage performance deviation, the gap between the current seepage state of the dam and the ideal state can be quantified, providing an objective evaluation basis. The preset error range is based on a large amount of historical data and engineering practice, possessing high scientific validity and accuracy, ensuring the reliability of the evaluation results. By periodically calculating the seepage performance deviation and comparing it with the preset error range, changes in the dam's seepage state can be detected in a timely manner, providing an effective early warning mechanism. The error range and model parameters are dynamically adjusted according to actual monitoring data, ensuring that the evaluation results always conform to the actual operating state of the dam. When the seepage performance deviation is within the error range, there is no need to start a complex seepage parameter inversion prediction model, saving computational resources and time. When the seepage performance deviation exceeds the error range, by starting the seepage parameter inversion prediction model, potential safety hazards can be identified more accurately, and corresponding prevention and remediation measures can be formulated. The model parameters can be flexibly selected and adjusted according to different deviation situations, improving the model's adaptability and prediction accuracy. It not only considers changes in a single parameter but also combines data from multiple dimensions such as water head, seepage flow, and pore water pressure to comprehensively reflect the seepage state of the dam.

[0072] In some embodiments of the present invention, the preset data acquisition time interval is dynamically adjusted according to the environmental conditions of the dam, seasonal changes, and extreme working conditions monitored in real time. Specifically:

[0073] The dam body is situated in a complex and diverse environment, including geological conditions (such as the properties of the soil and rock of the dam foundation and geological structure), hydrological conditions (such as changes in upstream and downstream water levels, rainfall, and evaporation), and meteorological conditions (such as temperature and wind speed). All of these environmental factors influence the seepage characteristics of the dam body. Different environmental conditions lead to variations in the seepage state of the dam body. For example, in areas with complex geological conditions and significant differences in soil and rock permeability, the seepage situation inside the dam body may be more complex and variable. In such cases, more frequent data collection is required, i.e., a shorter preset data collection interval is needed to promptly capture changes in the seepage state. Conversely, in areas with relatively simple and stable geological conditions, seepage changes are relatively slow, allowing for a more extended data collection interval. Similarly, when hydrological conditions change significantly, such as a sudden increase in rainfall or a rapid change in upstream and downstream water levels, parameters such as the dam body's seepage flow and head may change rapidly. To accurately grasp these changes, a shorter data collection interval is also necessary.

[0074] Seasonal changes can alter a range of factors related to dam seepage. For instance, during the rainy season, increased rainfall leads to greater head pressure on the dam, potentially resulting in a significant increase in seepage flow. Conversely, during the dry season, reduced rainfall leads to relatively stable seepage. Furthermore, seasonal changes can affect the physical properties of the dam foundation soil and rock; temperature variations can alter porosity and permeability. Given the impact of seasonal changes on dam seepage, it is reasonable to use different preset data collection intervals for different seasons. During the rainy season or periods of potential flooding, the data collection interval should be shortened to promptly monitor dam seepage and prevent leaks, landslides, and other safety incidents. Conversely, during the dry season, when dam seepage is relatively stable, the data collection interval can be appropriately extended, satisfying the need for monitoring dam seepage while reducing the workload of data collection and processing.

[0075] During real-time monitoring, various extreme conditions may be encountered, such as earthquakes, debris flows caused by heavy rainfall impacting the dam body, and localized damage to the dam body. These extreme conditions can severely affect the dam's structure and seepage characteristics, potentially causing drastic changes in the dam's seepage state within a short period. To cope with these extreme conditions and ensure the safety of the dam, when extreme conditions are detected in real-time monitoring, the preset data acquisition interval needs to be immediately adjusted and shortened. This allows for more intensive acquisition of seepage characteristic data from the dam body, timely capture of details of seepage changes under extreme conditions, and provides accurate data support for subsequent analysis and decision-making, enabling the implementation of effective emergency measures to ensure the safety of the dam.

[0076] In some embodiments of the present invention, the seepage parameter inversion prediction model achieves accurate prediction of the seepage characteristics of the dam body by combining the latest monitoring data (such as water head, seepage flow and pore water pressure) with optimized model parameters.

[0077] The seepage parameter inversion prediction model is mainly based on seepage mechanics theory. This theory describes the flow law of fluid in porous media (such as the dam body material of core rockfill dams), including basic principles such as Darcy's law. Darcy's law states that the seepage velocity is proportional to the hydraulic gradient, and the proportionality coefficient is the permeability coefficient. Through the above theory, the mathematical relationship between the seepage state of the dam body and the seepage parameters can be established, providing a theoretical basis for model construction.

[0078] The construction of the seepage parameter inversion prediction model relies on a large amount of monitoring data, including water head, seepage flow, and pore water pressure at a set location on the dam body. This data is not only used for model training, enabling it to learn the characteristic patterns under different seepage states, but also serves as input during model operation, driving the model to perform seepage parameter inversion prediction.

[0079] Specifically, the seepage parameter inversion prediction model adopts a numerical simulation-based model. Taking the finite element model as an example, the dam body is discretized into a finite number of elements, and seepage control equations, such as continuity equations and momentum equations, are established for each element and solved in combination with boundary conditions. During the solution process, based on the input seepage characteristic data and model parameters, seepage variables (such as head, velocity, etc.) within each element are calculated, thereby obtaining the seepage field distribution of the entire dam body. By continuously adjusting the model parameters, the simulation results are made to match the actual monitoring data as closely as possible, thereby realizing the inversion of seepage parameters.

[0080] On the other hand, seepage parameter inversion prediction models can also adopt machine learning-based models. Taking artificial neural network models as an example, they consist of an input layer, a hidden layer, and an output layer, with each layer connected by neurons. During training, a large amount of dam seepage feature data and its corresponding actual seepage parameters are used as training samples input into the model. By adjusting the weights and thresholds between neurons, the model can learn the complex nonlinear relationship between the input data and the seepage parameters. When a new set of dam seepage feature data is input, the model can output the corresponding predicted seepage parameter values ​​based on the learned relationship.

[0081] For models based on numerical simulation, the training process mainly involves adjusting model parameters to minimize the error between the simulation results and the actual monitoring data. Optimization algorithms, such as gradient descent and Newton's method, are typically used to find the optimal combination of model parameters. For models based on machine learning, the training process utilizes a large number of training samples to learn the model, adjusting the model's weights and parameters so that it can accurately classify or regress the input data to obtain accurate seepage parameter prediction results.

[0082] In practical engineering, seepage parameter inversion prediction models can receive seepage monitoring data of the dam body in real time and obtain the current seepage parameters of the dam body through inversion prediction. By comparing the above parameters with preset safety thresholds, if the seepage parameters are found to exceed the safe range, the model can issue an early warning signal in a timely manner, reminding the engineering personnel to take corresponding measures, such as checking whether there is leakage in the dam body and assessing the risk of seepage deformation, to ensure the safe operation of the dam body. The prediction results can also provide support for engineering decision-making. For example, in dam maintenance and reinforcement projects, based on the seepage parameters predicted by the model, the engineering personnel can evaluate the impact of different maintenance schemes on the seepage state of the dam body and select the optimal scheme. At the same time, in the design stage of new core rockfill dams, referring to the seepage parameter inversion prediction model results of similar projects can optimize the design parameters of the dam body and improve the seepage prevention performance and stability of the dam body.

[0083] like Figure 2 As shown, the seepage parameter inversion and prediction system for core-wall rockfill dams of the present invention specifically includes the following modules;

[0084] The target score acquisition module is used to acquire the target seepage performance score of the dam body;

[0085] The data acquisition module is used to collect operational data of the core rockfill dam according to a preset acquisition time interval, and obtain a set of seepage characteristic data of the dam body;

[0086] The real-time analysis module is used to analyze and obtain a real-time seepage performance score of the dam body based on multiple sets of acquired dam body seepage characteristic data;

[0087] The decision module is used to determine whether seepage inversion prediction is needed based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body; and when seepage inversion prediction is needed, to determine the model parameters of the preset seepage parameter inversion prediction model.

[0088] The inversion prediction module is used to input the latest acquired dam seepage characteristic data set into the seepage parameter inversion prediction model with confirmed model parameters to obtain the inversion prediction results of seepage parameters for core rockfill dams.

[0089] In this embodiment, firstly, the data acquisition module collects data at preset intervals, and the real-time analysis module derives a real-time seepage performance score based on this data. Compared to traditional methods that rely on historical data, this approach can promptly detect seepage anomalies and respond quickly even under extreme conditions or localized damage, avoiding prediction lag. Secondly, the decision-making module improves prediction accuracy and resource utilization efficiency. By comparing real-time and target seepage performance scores, it performs inversion prediction and determines model parameters only when necessary. This avoids unnecessary waste of computational resources and allows for model optimization based on actual needs, making the inversion prediction more closely match the actual seepage situation of the dam body and improving prediction accuracy. Finally, the inversion prediction module ensures dam safety. By inputting the latest data into a model with adapted parameters, the obtained inversion prediction results provide a scientific basis for dam seepage safety assessment, helping staff to promptly grasp changes in dam seepage characteristics, identify potential safety hazards in advance, and take effective measures to ensure the structural stability and long-term service performance of the dam.

[0090] The various variations and specific embodiments of the seepage parameter inversion prediction method for core-rockfill dams in the aforementioned Embodiment 1 are also applicable to the seepage parameter inversion prediction system for core-rockfill dams in this embodiment. Through the aforementioned detailed description of the seepage parameter inversion prediction method for core-rockfill dams, those skilled in the art can clearly understand the implementation method of the seepage parameter inversion prediction system for core-rockfill dams in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0091] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting the seepage parameters of a core rock-fill dam by inversion, characterized in that, The method comprises the following steps: obtaining a target seepage performance score of the dam body; collecting operation data of the core rock-fill dam according to a preset collection time interval to obtain a plurality of dam body seepage characteristic data sets; based on the obtained plurality of dam body seepage characteristic data sets, analyzing to obtain a real-time seepage performance score of the dam body; based on the real-time seepage performance score of the dam body and the target seepage performance score of the dam body, determining whether seepage inversion prediction is needed; and when seepage inversion prediction is needed, determining the model parameters of a preset seepage parameter inversion prediction model; comprising: calculating a seepage performance deviation amount according to the real-time seepage performance score of the dam body and the target seepage performance score of the dam body; comparing the seepage performance deviation amount with a preset error range: if the seepage performance deviation amount is within the preset error range, it indicates that the current seepage state is close to the ideal state, and no inversion prediction is needed; if the seepage performance deviation amount exceeds the preset error range, seepage inversion prediction is needed, and the model parameters of the preset seepage parameter inversion prediction model are determined according to the seepage performance deviation amount; inputting the latest obtained dam body seepage characteristic data set into the seepage parameter inversion prediction model with the confirmed model parameters to obtain an inversion prediction result of the core rock-fill dam seepage parameter; the determination of the model parameters of the preset seepage parameter inversion prediction model according to the seepage performance deviation amount comprises: based on the seepage performance deviation amount, analyzing the change relationship of the water flow velocity or pressure distribution inside the dam body in combination with the seepage physical mechanism; when the seepage performance deviation amount exceeds the preset error range and shows an upward trend, adjusting the parameters representing the water flow velocity or pressure distribution in the seepage parameter inversion prediction model; optimizing the model parameters by using a mathematical optimization algorithm, comprising: constructing an objective function, which is based on the seepage performance deviation amount and the observation data in the dam body seepage characteristic data set, and represents the difference degree between the model prediction value and the actual observation value; setting the initial value of the model parameters; iteratively adjusting the model parameters to minimize the objective function or to meet the convergence condition.

2. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 1, characterized in that, The method for obtaining the target seepage performance score of the dam body comprises: obtaining the design standard of the core rock-fill dam to be inverted, identifying and extracting the seepage parameter data to obtain a target seepage parameter set; performing seepage performance analysis on the target seepage parameter set to obtain the target seepage performance score of the dam body.

3. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 2, characterized in that, The target seepage parameter set comprises a target permeability coefficient and a target hydraulic slope.

4. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 3, characterized in that, The dam body seepage characteristic data set comprises the water head, seepage flow and pore water pressure at a set position of the dam body.

5. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 4, characterized in that, Based on the obtained plurality of dam body seepage characteristic data sets, the real-time seepage performance score of the dam body is analyzed and obtained, comprising: sorting and aligning a set number of dam body seepage characteristic data sets in the order of collection time stamps to obtain a dam body seepage characteristic matrix; inputting the dam body seepage environment characteristic matrix into a pre-constructed seepage performance evaluation model to obtain the real-time seepage performance score of the dam body.

6. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 4, characterized in that, The seepage performance evaluation model is constructed by using a multi-layer perception architecture, the number of nodes of an input layer is equal to the dimension of a seepage feature matrix, a hidden layer contains neurons with a ReLU activation function, and an output layer generates a normalized score value by using a Sigmoid function.

7. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 1, characterized in that, The preset collection time interval is dynamically adjusted according to the environmental conditions of the dam body, seasonal changes and real-time monitored extreme working conditions.

8. The heart wall rock-fill dam seepage parameter inversion prediction method according to claim 1, characterized in that, The preset error range is determined according to historical operation data, design standards and safety margins of the core rock-fill dam, and is periodically optimized and adjusted according to the latest obtained operation data.

9. A system for predicting the seepage parameters of a core wall rockfill dam, the system being applied to the method for predicting the seepage parameters of a core wall rockfill dam according to claim 1, characterized in that, The system comprises: a target score acquisition module configured to acquire a target seepage performance score of the dam body; a data collection module configured to collect operation data of the core rock-fill dam according to a preset collection time interval, and obtain a set of seepage feature data of the dam body; a real-time analysis module configured to analyze and obtain a real-time seepage performance score of the dam body based on the obtained set of seepage feature data of the dam body; a decision module configured to determine whether seepage inversion prediction is needed according to the real-time seepage performance score of the dam body and the target seepage performance score of the dam body, and determine model parameters of a preset seepage parameter inversion prediction model when seepage inversion prediction is needed; an inversion prediction module configured to input the latest obtained set of seepage feature data of the dam body into the seepage parameter inversion prediction model with the confirmed model parameters, and obtain an inversion prediction result of seepage parameters of the core rock-fill dam.

Images