Earth and rockfill dam three-dimensional seepage simulation and parameter inversion analysis method and system based on finite element method

Abstract

The invention discloses an earth and rockfill dam three-dimensional seepage simulation and parameter inversion analysis method and system based on a finite element method, and relates to the technical field of seepage simulation and parameter inversion analysis of geotechnical engineering. Comprising the following steps: deriving an earth and rockfill dam unstable unsaturated seepage control differential equation based on a Darwest unsaturated seepage theory, solving the equation based on a finite element method, and developing a software platform HYDRO-GEO / S based on a Matlab programming language; establishing an earth and rockfill dam three-dimensional seepage finite element model; performing three-dimensional seepage calculation and seepage safety analysis on the earth and rockfill dam; and constructing a parameter inversion framework to obtain the hydraulic parameters of the material. According to the method, an unsaturated seepage theory and a finite element method are coupled, an ICSSA-RBF inversion framework is constructed, an HYDRO-GEO / S software platform is developed, and in combination with instance analysis, accurate simulation and safety evaluation of earth and rockfill dam seepage and efficient inversion of seepage parameters are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of seepage simulation and parameter inversion analysis of geotechnical engineering, in particular to a three-dimensional seepage simulation and parameter inversion analysis method and system for earth-rock dams based on the finite element method. BACKGROUND

[0002] The safety performance of the anti-seepage structure of a high earth-rock dam directly affects the water storage capacity of the reservoir and the safety of the dam body. Abnormal seepage problems not only cause abnormal distribution of water pressure in the dam body, but if further deteriorate, may cause dam instability and even catastrophic consequences such as dam collapse, and cause unacceptable secondary disasters. How to ensure the seepage safety of high earth-rock dams has always been an important and key issue in the field of hydraulic geotechnical engineering.

[0003] Seepage numerical analysis is an important means to ensure the safety of high earth-rock dam projects, optimize design, improve economy, and promote scientific research. However, traditional two-dimensional seepage analysis cannot accurately reflect the real complex seepage behavior of high earth-rock dams in three-dimensional space. Carrying out three-dimensional seepage analysis of high earth-rock dams can more realistically simulate the seepage behavior of the dam body and provide a scientific basis for optimizing anti-seepage design and assessing dam safety. However, the strong unsaturated characteristics of high earth-rock dams pose challenges to the stability of numerical analysis. At the same time, hydraulic parameters such as permeability coefficient and unsaturated parameters are the core data of high earth-rock dam seepage numerical analysis, which directly affect the numerical simulation results of the seepage flow field. However, these parameters are often difficult to accurately obtain through experiments. Through inversion analysis based on monitoring data such as pressure head and seepage flow, the hydraulic parameters of the dam body and foundation can be optimized and adjusted, thereby improving the accuracy of seepage analysis and providing a scientific basis for engineering design. Therefore, developing an efficient hydraulic parameter inversion method is also very important for the seepage safety of high earth-rock dams. SUMMARY

[0004] To solve the above technical problems, a three-dimensional seepage simulation and parameter inversion analysis method for earth-rock dams based on the finite element method is provided, which solves the above problems.

[0005] To achieve the above purposes, the technical scheme adopted by the present application is as follows:

[0006] A three-dimensional seepage simulation and parameter inversion analysis method for earth-rock dams based on the finite element method, comprising:

[0007] Develop a three-dimensional seepage calculation and analysis software platform, including deriving the non-steady unsaturated seepage control differential equation of the earth-rock dam based on the Darcy unsaturated seepage theory, solving the equation based on the finite element method, and developing the software platform HYDRO-GEO / S based on the Matlab program language;

[0008] A three-dimensional seepage finite element model of earth-rock dam is established, including establishing a three-dimensional entity model and dividing finite element grid according to the data of earth-rock dam, determining initial conditions and boundary conditions of seepage analysis and determining material parameters;

[0009] Three-dimensional seepage calculation and seepage safety analysis of earth-rock dam are carried out, including executing calculation of seepage calculation program and finite element model to obtain results, and evaluating the safety of core wall, curtain grouting and the like based on comparison of critical hydraulic gradient with calculation results of hydraulic gradient;

[0010] A parameter inversion framework is constructed to obtain hydraulic parameters of materials, including building a sparrow search algorithm-radial basis function neural network SSA-RBF framework, optimizing to ICSSA-RBF framework, and combining monitoring data and finite element calculation to invert hydraulic parameters.

[0011] Preferably, the development of the three-dimensional seepage calculation and analysis software platform comprises deriving unsteady unsaturated seepage control differential equations of earth-rock dam based on Darcy unsaturated seepage theory, solving the equations based on finite element method, and developing software platform HYDRO-GEO / S based on Matlab program language.

[0012] Based on the principle of mass conservation, the difference between the water inflow and outflow rates of the differential element is equal to the change rate of water content, and the continuity equation of unsteady unsaturated seepage can be obtained:

[0013]

[0014] In the formula, ρ is the density of water, v i is the Darcy velocity, n is the porosity, S w is the saturation, nS w = θ is the water content.

[0015] Unlike saturated seepage, unsaturated calculation needs to consider the change process of element water content, and uses pressure head h p = p / ρg and volume water content θ = nS w to replace total head h and saturation S w , and further obtains the control differential equation of unsteady unsaturated seepage:

[0016]

[0017] In the formula, is the water capacity, which represents the change value of unit volume water content caused by unit pressure head change, and is 0 in the saturated area, β is 1 and 0 in the saturated area and the unsaturated area respectively, and is often taken as 0 in the saturated area in water conservancy engineering, then pressure head is used to replace, is the ratio of relative permeability, which is 1 in the saturated area, and needs to be fitted according to the soil-water characteristic curve in the unsaturated area.

[0018] Preferably, the development of the three-dimensional seepage calculation and analysis software platform comprises deriving the non-steady unsaturated seepage control differential equation of the earth-rock dam based on the Darcy non-saturation seepage theory, solving the equation based on the finite element method, and developing the software platform HYDRO-GEO / S based on the Matlab program language, and the software platform further comprises:

[0019] The matrix format of the three-dimensional finite element seepage integral equation is derived as follows:

[0020]

[0021] In the formula, {f} is the known water head of the initial boundary, [K] is the seepage matrix, [S] is the matrix reflecting the soil water compression capacity and the infiltration evaporation capacity, and [P] is the flow at the flow boundary. The time term involved in the matrix format of the three-dimensional finite element seepage integral equation is taken as an implicit finite difference format:

[0022]

[0023] Without considering the influence of soil water compression, it is further obtained that:

[0024]

[0025] In the initial steady flow calculation, it is further degenerated as:

[0026] [K]{h}={f}

[0027] On the basis of the matrix format of the derived three-dimensional finite element seepage integral equation, the convergence of the seepage matrix in the earth-rock dam is improved by combining the matrix scaling, nested partitioning and ordering, and the LU decomposition preprocessing method, and then the iterative solution of the unsaturated seepage in the earth-rock dam is completed by using the conjugate gradient solving method.

[0028] Preferably, the development of the three-dimensional seepage calculation and analysis software platform HYDRO-GEO / S based on the Matlab program language specifically comprises:

[0029] In the preprocessing stage, the EditField, DropDown and Slider controls of the MATLAB AppDesigner are used to realize the definition of the material permeability coefficient, porosity, boundary condition, seepage calculation type, element type and time step, and the UIAxes component is used to build a soil water characteristic curve visualization window, and the boundary condition parameter input is completed through a preprocessing file;

[0030] In the interactive logic, the Button callback function is used to trigger different calculation processes, including selecting the saturated and unsaturated seepage matrix assembly mode, and the direct method and the preprocessing conjugate gradient method for solving large-scale linear equations;

[0031] In the post-processing stage, statistical analysis functions are integrated to support the export of calculation results of hydraulic gradient and pressure head. The App is compiled into an independent executable file using MATLAB Compiler and packaged into an installation package using MLAppinstall to ensure stable operation on engineering terminals without a MATLAB environment.

[0032] Preferably, the establishment of the three-dimensional seepage finite element model of the earth-rock dam includes, based on the earth-rock dam data, establishing a three-dimensional solid model and dividing it into finite element meshes, determining the initial and boundary conditions for seepage analysis, and determining the material parameters, specifically including:

[0033] Using Ansys SpaceClaim software, based on geological exploration data (including stratigraphic distribution and lithological parameters) and design data (including dam height and dam axis dimensions), a three-dimensional solid model of the earth-rock dam was completed through operations such as sweeping, cutting, and combining.

[0034] Hypermesh software was used to generate a three-dimensional finite element mesh based on the model size and accuracy requirements.

[0035] Preferably, the establishment of the three-dimensional seepage finite element model of the earth-rock dam includes establishing a three-dimensional solid model based on the earth-rock dam data and dividing it into finite element meshes, determining the initial and boundary conditions for seepage analysis, and determining the material parameters, and further includes:

[0036] The specific method for determining the initial and boundary conditions for seepage analysis of earth-rock dams is as follows:

[0037] To determine the first type of boundary condition, considering the unsteady seepage boundary condition related to time t, the expression is:

[0038]

[0039] S204: Determine the second type of boundary conditions, the expression is:

[0040] When considering anisotropy, the expression is:

[0041]

[0042] In the formula, q represents the flow rate transmitted per unit area through the boundary, equivalent to v. n , l x ,l y ,l z Let n be the direction cosine between the outward normal n and the coordinate;

[0043] The third type of boundary condition is determined, and the linear relationship between the head difference inside and outside the aquifer boundary and the exchange flow is established. The expression is as follows:

[0044]

[0045] In the formula, α is a positive constant, and both α and β are known values ​​at each point of this type of boundary;

[0046] The initial conditions were determined, and the first type of boundary conditions were used as the initial state for seepage analysis.

[0047] Preferably, the establishment of the three-dimensional seepage finite element model of the earth-rock dam includes establishing a three-dimensional solid model based on the earth-rock dam data and dividing it into finite element meshes, determining the initial and boundary conditions for seepage analysis, and determining the material parameters, and further includes:

[0048] The specific method for determining the material parameters for seepage analysis of earth-rock dams is as follows:

[0049] The finite element model material is divided into saturated and unsaturated regions;

[0050] For materials in the unsaturated region, the relationship between pore water pressure and saturation is described using the soil-water characteristic curve (SWCC curve), fitted by the Van Genuchten model, with the expression as follows:

[0051]

[0052] In the formula, Let be a function of the volumetric water content of the soil. For matrix suction, S e The degree of saturation is given by 'a', which represents the suction force corresponding to the inflection point of the soil-water characteristic curve, and 'n' represents the slope at the inflection point of the soil-water characteristic curve. 'm' = 1 - 1 / n.

[0053] Preferably, the three-dimensional seepage calculation and seepage safety analysis of the earth-rock dam includes executing calculations using a seepage calculation program and a finite element model to obtain results, and assessing the safety of the core wall, curtain grouting, and other components based on the hydraulic gradient calculation results and the critical hydraulic gradient. Specifically, this includes:

[0054] Export the generated 3D finite element mesh information as a text file containing element node coordinates. The file format includes element number, node number, and node coordinates.

[0055] The mesh information text file is processed, and the defined boundary conditions and material parameters are imported through the software platform;

[0056] Perform three-dimensional seepage calculations in the HYDRO-GEO / S platform and export the hydraulic gradient and pressure head distribution results.

[0057] Seepage safety analysis involves extracting the calculated hydraulic gradient values ​​of key components such as the core wall and curtain grouting, and comparing them with the critical hydraulic gradient. If the calculated hydraulic gradient value is less than or equal to the critical hydraulic gradient, the seepage prevention material is deemed safe. If the calculated hydraulic gradient value is greater than the critical hydraulic gradient, there is a risk of seepage failure.

[0058] Preferably, the construction of the parameter inversion framework to obtain the hydraulic parameters of the material includes building a Sparrow Search Algorithm-Radial Basis Function Neural Network (SSA-RBF) framework, optimizing it into an ICSSA-RBF framework, and combining monitoring data with finite element calculation to invert the hydraulic parameters, specifically including:

[0059] Based on the radial basis function neural network framework and employing the sparrow search algorithm, the specific method for constructing the sparrow search algorithm-radial basis function neural network parameter inversion framework is as follows:

[0060] The radial basis function (RBF) neural network framework comprises an input layer, hidden layers, and an output layer. Based on the input vector K of the input layer, the hidden layer space constructed using radial basis functions performs a nonlinear transformation on the input vector K, increasing its dimensionality to make it linearly separable in a high-dimensional space, thus linearly outputting the desired result H. This invention uses the Gaussian function, the most commonly used basis function in RBF neural networks, as the basis function, with the following format:

[0061]

[0062] In the formula, K is an m-dimensional input vector; c i It is the center point of the i-th basis function with the same dimension as K; ||Kc i || is the vector norm, representing the relationship between K and c. i Euclidean distance; σ i The variance of the Gaussian function is represented by the fact that the dimension I of the center point is greater than the dimension m of the input vector, thus completing the nonlinear mapping of the input parameters from low dimension to high dimension.

[0063] The following formula is used to obtain the result required for the linear output of the hidden layer:

[0064]

[0065] In the formula, W in is the weight coefficient between output node n and the i-th center point, and N is the number of output nodes n;

[0066] The sparrow search algorithm is used to find the optimal parameters of the radial basis function neural network surrogate model. Based on the defined sparrow population size S, the position X(X1…X2) of each sparrow is initialized. i …X q ), and calculate the fitness f(X1…X i …X qBased on the sparrow population size S, the discoverers searching for food are defined at a fixed ratio (10%–20%), and the positions of the discoverers are updated at each iteration using the following formula:

[0067]

[0068] In the formula, Let λ be the position of sparrow i at dimension j during the t-th iteration, and λ be a random number in the range (0,1]. max t is the maximum number of iterations; Q is a random number that follows a normal distribution; R is the warning value for the sparrow search range, which ranges from [0,1]; S T is the safe threshold for sparrows to forage, which ranges from [0.5,1]. L is a 1×q matrix with all values ​​of 1. When the warning value is less than the safe threshold, sparrows are in a state of having no natural enemies and can search for food widely. Otherwise, in order to avoid danger, sparrows will move randomly in a normal distribution and forage in a small area.

[0069] The followers of a sparrow population, excluding the discoverer, constitute the main group of the population. They obtain food based on the discoverer's location, and there is a mutual surveillance relationship among the followers. Followers will always tend to approach the location with the highest food availability, and their locations are updated according to the following strategy:

[0070]

[0071] In the formula, It is the most unfavorable position in the t-th iteration; It is the optimal position in the (t+1)th iteration; A + =A T (AA T ) -1 It is the Moore-Penrose generalized inverse of matrix A, where A is a matrix consisting of -1 and 1. When i > n / 2, it means that the follower has difficulty obtaining food and needs to search for other locations to forage; otherwise, it means that the follower will compete with the discoverer for food and may replace the discoverer's position.

[0072] Within a sparrow population, 10% to 20% are vigilant individuals. When they detect danger, they promptly adjust their positions to avoid it. The vigilant individuals' position update strategy is as follows:

[0073]

[0074] In the formula, It is the optimal position in the t-th iteration; f i This is the current fitness value of the sparrow; f g It is the fitness value of the current global best position; f wis the fitness value of the current globally unfavorable position; β is a random number that follows a standard normal distribution; η is a random number in the range [0,1]; ε is a constant to avoid a denominator of 0, f i >f g This means that sparrows are on the edge of the population and are more vulnerable to predators; f i =f g When this happens, it indicates that the sparrows in the middle of the population are aware of the threat from predators and will move to avoid the risk.

[0075] The specific method for constructing the optimized sparrow search-radial basis neural network inversion framework by introducing a reverse learning strategy, Lévy flight random walk, and Logistic-Tent mapping and variable step size search optimization strategy is as follows:

[0076] A reverse learning strategy is introduced. Based on the current optimal solution, the corresponding reverse solution is obtained, the reverse solution is compared and evaluated with the optimal solution, and the current optimal solution is updated to expand the optimization ability and scope of each individual in the population. The strategy is as follows:

[0077]

[0078] In the formula, ub and lb are the proposed upper and lower boundary values, respectively; yes The inverse solution; r is a 1×q random matrix uniformly distributed in the range (0,1); b=(Tt / T) t These are control parameters related to information exchange;

[0079] The Levy flight method is introduced to address the problem of sparrows getting trapped in local optima during search. The sparrow random walk strategy after the Levy flight perturbation is as follows:

[0080]

[0081]

[0082] To address the search objectives before and after the initial search, a position update weight is introduced, and the discoverer's position update strategy is as follows:

[0083]

[0084] In the formula, ω t ω0 is the weight of the current iteration step t; ω0 is a constant greater than 1; t0 is the number of iterations at which the local search capability is enhanced.

[0085] The initial sparrow samples were optimized using Logistic-Tent mapping to make their distribution more uniform. The mapping relationship of the initial samples was as follows:

[0086]

[0087] In the formula, r is a random number in (0,4). Based on the updated sparrow search method described above, the search for the optimal radial basis parameters is completed quickly and efficiently, and finally the optimized sparrow search-radial basis neural network inversion framework is constructed.

[0088] The specific method for obtaining the hydraulic parameters of materials by combining monitoring data and finite element calculation inversion is as follows:

[0089] The evaluation indicators for inversion analysis are determined, and monitoring data for these indicators are obtained through exploration. Water head is the most easily measurable hydraulic parameter of the dam site's natural foundation and dam structure, and it is also a hydraulic indicator that can be directly calculated using seepage finite element analysis. Therefore, the objective function of the radial basis function is established based on the water head.

[0090]

[0091] In the formula, H n It is the calculated pressure head value at the target location n; H n ′ represents the inversion head value at the target location n; This is the weight at the target position n, reflecting the importance of the data at that point within the overall data. When the importance of nodes is the same, the weight is taken as the highest value.

[0092] Based on the finite element model established in S4-S6, the developed seepage calculation and analysis platform is called within the ICSSA-RBF framework to perform finite seepage analysis and complete the inversion of hydraulic parameters to obtain the required hydraulic parameters.

[0093] A three-dimensional seepage simulation and parameter inversion analysis system for earth-rock dams based on the finite element method includes:

[0094] Software platform development module, finite element model construction module, seepage calculation and safety analysis module, parameter inversion module;

[0095] The software platform development module is used to develop a three-dimensional seepage calculation and analysis software platform HYDRO-GEO / S with pre-processing, calculation and post-processing functions based on Darcy's unsaturated seepage theory, finite element method and Matlab language.

[0096] The finite element model building module is used to complete three-dimensional solid modeling and mesh generation based on earth-rock dam data, define the initial conditions, boundary conditions and material parameters for seepage analysis, and include soil-water characteristic curves fitted by the VanGenuchten model.

[0097] The seepage calculation and safety analysis module is used to import the finite element model into the HYDRO-GEO / S platform to perform three-dimensional seepage calculation, export the hydraulic gradient results and compare them with the critical hydraulic gradient to evaluate the seepage safety of the core wall and curtain grouting.

[0098] The parameter inversion module is used to build an SSA-RBF framework and optimize it into an ICSSA-RBF framework. Combined with head monitoring data and finite element calculations, it inverts and obtains the hydraulic parameters of earth-rock dam materials.

[0099] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0100] This invention proposes a differential equation for the control of unsteady unsaturated seepage in earth-rock dams based on Darcy's unsaturated seepage theory, and uses the finite element method to iteratively solve the seepage problem. Furthermore, based on the Matlab programming language, a three-dimensional seepage analysis software platform, HYDRO-GEO / S, was independently developed. For the hydraulic parameter inversion problem, a sparrow search algorithm based on a radial basis function neural network framework was established, incorporating a back-learning strategy and Lévy flight random walk. This optimized sparrow search algorithm-radial basis function neural network parameter inversion framework provides a powerful solution for three-dimensional seepage analysis and hydraulic parameter acquisition in complex earth-rock dam projects, solving the challenges of hydraulic parameter inversion and seepage safety analysis of seepage barriers in high earth-rock dams. Attached Figure Description

[0101] Figure 1 The flowchart shows the three-dimensional seepage simulation and parameter inversion analysis method for earth-rock dams based on the finite element method.

[0102] Figure 2 This is a screenshot of the HYDRO-GEO / S software platform interface.

[0103] Figure 3 A schematic diagram of a three-dimensional finite element model of a 300m high core wall rockfill dam;

[0104] Figure 4 This is the maximum cross-sectional view of the earth-rock dam model;

[0105] Figure 5 This is a schematic diagram of a three-dimensional finite element model of the seepage prevention system;

[0106] Figure 6 The soil-water characteristic curves and hydraulic conduction curves of the core wall material and the filter material; Figure 6 (a) is a water content curve. Figure 6 (b) is a hydraulic transmission curve.

[0107] Figure 7 This is a diagram showing the hydraulic gradient distribution at the maximum cross-section of the core wall.

[0108] Figure 8 Hydraulic gradient distribution diagram of curtain grouting;

[0109] Figure 9 This is a schematic diagram showing the distribution of dam foundation materials;

[0110] Figure 10 This is a schematic diagram of the finite element model of the dam foundation;

[0111] Figure 11 A schematic diagram showing the absolute head difference under different inversion analysis strategies;

[0112] Figure 12 A diagram showing the error comparison of different inversion analysis strategies;

[0113] Figure 13 This is a diagram showing the initial flow field pressure and head distribution obtained from the inversion analysis.

[0114] Figure 14 A schematic diagram of the free water surface distribution obtained for inversion analysis. Detailed Implementation

[0115] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0116] Example 1

[0117] This invention provides a method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method. Specifically, the basic process can be as follows: Figure 1 As shown, it includes:

[0118] Step S1 derives the unsteady unsaturated seepage control differential equation for earth-rock dams based on Darcy's unsaturated seepage theory. The final seepage control differential equation is:

[0119]

[0120] In step S2, the unsteady unsaturated seepage control differential equation is solved based on the finite element method. First, according to equation (1), the matrix format of the three-dimensional finite element seepage global equation is obtained as follows:

[0121]

[0122] Based on equation (2), a finite difference scheme is adopted for the time term, and the convergence of the permeability matrix in the earth-rock dam is further improved by combining matrix scaling, nested partitioning sorting, LU decomposition and other preprocessing methods. Then, the conjugate gradient solution method is used to complete the iterative solution of the unsaturated seepage of the earth-rock dam.

[0123] In step S3, a three-dimensional seepage calculation and analysis software platform is developed based on the Matlab programming language. The platform interface is as follows: Figure 2 As shown in the diagram, this software platform allows for the definition of key calculation parameters such as material permeability coefficient, porosity, boundary conditions, seepage calculation type, element type, and time step during the preprocessing stage. In terms of interactive logic, different calculation processes can be triggered, including saturated and unsaturated permeability matrix assembly modes, and efficient methods for solving large-scale linear equation systems, such as the direct method and the preprocessing conjugate gradient method, to complete the three-dimensional seepage calculation. In the post-processing stage, integrated statistical analysis functions allow results to be exported to Excel and other formats. Finally, the App is compiled into a standalone executable file using MATLAB Compiler and packaged into an installation package using MLAppinstall to ensure stable operation on engineering terminals without a MATLAB environment.

[0124] In step S4, a three-dimensional solid model of the earth-rock dam is established based on the earth-rock dam model data, and the finite element mesh of the model is generated. The three-dimensional high-core rockfill dam used in this invention example has a height of 300m. The natural strata include a weakly unloaded zone, a weakly weathered upper zone, a weakly weathered lower zone, and fresh rock mass. The depth of the relatively impermeable fresh rock mass is taken as 400m, the length on the left and right sides is taken as 500m, and the length on the upstream and downstream sides is taken as 900m. The number of elements in the finite element model is 451,890, and the number of nodes is 470,137. Among them, the number of elements and nodes in the dam body are 153,620 and 168,537, respectively. The overall three-dimensional finite element model is shown below. Figure 3 As shown. The maximum cross-sectional view of the earth-rock dam model is as follows. Figure 4 The three-dimensional finite element model of the seepage prevention system is as follows: Figure 5 As shown, the curtain grouting is inserted into the fresh bedrock for 10m, with a maximum depth of 100m.

[0125] Step S5 determines the initial and boundary conditions for the seepage analysis of the earth-rock dam. The dam crest elevation is 300m, and the upstream water level is taken as 290m for the calculation. After the water level is determined, the boundary conditions for the calculation are as follows: ① Known head boundary, which includes the reservoir bank below the upstream water level of the dam site, the upstream slope of the dam body, and the river channel; ② Seepage boundary, which includes the left and right bank slopes above the upstream water level of the reservoir area, the downstream slope of the dam body, the dam crest, and the downstream side of the model; ③ Impermeable boundary, which includes the upstream side of the model, the boundaries of the left and right banks, and the boundary of the bedrock bottom of the model. In this example, only the steady-state seepage calculation analysis of the earth-rock dam is performed, so there is no need to set initial conditions.

[0126] Step S6 determines the material parameters for seepage analysis of the earth-rock dam. The dam body materials are divided into five types: rockfill, transition material, filter layer 1, filter layer 2, and core wall. The dam foundation materials include four types: weak unloading zone, weakly weathered upper zone, weakly weathered lower zone, and fresh rock mass. The clay seepage control material for high-clay rockfill dams is crucial for ensuring dam safety due to its extremely strong hygroscopic properties and excellent seepage control performance. Therefore, unsaturated seepage analysis is needed to study the seepage characteristics of high earth-rock dams. Considering that the filter materials on both sides of the core wall serve as auxiliary seepage control materials, their unsaturated seepage characteristics are equally important. Soil-water characteristic curves and hydraulic conduction curves for the core wall material and filter materials are derived, such as... Figure 6 As shown. Furthermore, Table 1 provides the permeability coefficients for all materials;

[0127] Table 1

[0128]

[0129] In step S7, the three-dimensional seepage calculation of the earth-rock dam is completed, and the calculation results are exported. The finite element model with boundary condition processing completed in S5 and the material parameters determined in S6 are input into the seepage calculation platform developed in S3. The seepage calculation is completed using the software platform, and the calculation results are exported.

[0130] Step S8 completes the seepage safety analysis of the earth-rock dam. In this invention example, the seepage safety analysis is mainly performed on the core wall and curtain wall, which are important seepage prevention components of the earth-rock dam.

[0131] (1) Safety analysis of seepage in the core wall: based on the critical hydraulic gradient J cr To establish a seepage stability standard, the mean hydraulic gradient is used to provide feedback on the overall seepage safety status of the core wall. Figure 7 The distribution of the hydraulic gradient at the maximum cross-section of the core wall of the earth-rock dam in this embodiment of the invention is shown. Under the condition of the highest water level, the maximum hydraulic gradient in the saturated region of the core wall is 2.65, which occurs near the overflow point of the core wall, and the mean hydraulic gradient in the saturated region of the core wall is 2.13. The calculation results show that, under the evaluation criterion of a critical hydraulic gradient of 4.0, seepage failure will not occur in the core wall of this earth-rock dam.

[0132] (2) Safety analysis of curtain seepage: Figure 8 The hydraulic gradient distribution of the curtain grouting within the dam foundation is shown, with the largest hydraulic gradient value (23.5) observed in the surface region at the bottom of the valley, where the water pressure is greatest. Although this value is much larger than that of porous media, the critical failure gradient value of concrete typically exceeds 100, thus indicating that it remains in a very safe state.

[0133] Steps S9 and S10 complete the construction of the optimized sparrow search-radial basis neural network inversion framework;

[0134] Step S11, based on the parameter inversion framework established in S9 and S10, combines the finite element model and monitoring data to invert and obtain the hydraulic parameters of the material. The mechanical properties of dam materials can be easily measured through experimental methods. However, the foundation area involved in earth-rock dam reservoirs is extensive and the geological conditions are complex, making it difficult to accurately obtain the hydraulic parameters of natural foundation materials through experimental methods. Here, material parameter inversion is performed on the foundation of a reinforced concrete face rockfill dam used in a certain water conservancy project. Similar to the process described in steps S4 and S5 above, a finite element model of the natural dam foundation is established, see... Figure 9 and Figure 10 ;

[0135] The groundwater depth on the left bank of the dam site ranges from 18m to 82.5m, while that on the right bank ranges from 22m to 75m. Table 2 shows the groundwater level statistics from the boreholes in the dam site area. There were 26 groundwater level exploration points, and the free water surface elevation at each point was measured. Table 3 shows the permeability parameters of the weathered rock mass in each borehole in the dam site area, based on the exploration test data.

[0136] Table 2

[0137]

[0138] Table 3

[0139]

[0140] For the four main rock masses present in the dam foundation, 108 sets of initial permeability coefficient parameter combinations were taken, covering all representative values ​​for each rock mass in Table 4, and the head calculation results for the exploration points in Table 2 were obtained for each set of parameters. Considering the large range of values ​​for completely weathered surface material and its relative sensitivity to the location and distribution of the free water surface, four representative permeability coefficient values ​​were proposed for this area, ranging from the upper limit (most unfavorable permeability coefficient 1×10⁻⁵ m / s) to the lower limit (permeability coefficient 5×10⁻⁵ m / s). For strongly weathered and weakly weathered materials, upper, median, and lower limits were proposed, respectively. Given that fresh rock masses are basically saturated, three representative values ​​were also proposed. For the above 108 sets of data, during RBF neural network training, 80 sets containing the upper and lower limits of all materials were used as the training sample set, and the remaining 28 sets were used as the validation sample set. To verify the parameter inversion method proposed in this invention, the results of RBF, SSA-RBF, and ICSSA-RBF training models were compared.

[0141] Table 4

[0142]

[0143] The inversion results for different radial basis function neural network search strategies are given in Table 5. Figure 11A comparison of the optimal calculation results for each exploration point is presented. The results show that the optimized radial basis function neural network (RBN) of the sparrow search strategy is closest to the actual water surface elevation at 18 exploration points, including ZK101 and ZK102, while the original RBN has the worst inversion effect, with only 3 exploration points being closest to the actual water surface. Due to the limitations of the original sparrow search strategy, such as local optima, 5 exploration points are closest to the actual water surface without optimization. According to the optimal scheme of the ICSSA RBN search proposed in this invention, the exploration point ZK117 has the largest error, with the inverted free water surface elevation differing from the exploration elevation by 1.44m. The errors at the other points are generally within 1m. Considering the large number of exploration points used for inversion, the above errors are acceptable.

[0144] Table 5

[0145]

[0146]

[0147] Figure 12 The calculation results of the relative average head difference Z under different strategies are presented. The results show that the original RBF optimal parameter search performs the worst, resulting in low accuracy of the inversion results. However, after the initial sparrow optimization search, the inversion performance is significantly improved. Based on SSA-RBF, the method proposed in this invention further reduces the inversion error from 4.7‰ to 3.3‰, improving the inversion accuracy by 29.8%. Furthermore, it can be observed that even with the significantly better optimized sparrow search method, the inversion results still have some errors. This is due to uncertainties such as the difference between the numerical model and real engineering conditions, and the heterogeneity of materials.

[0148] Based on the hydraulic parameters obtained from the inversion analysis, the overall pressure head distribution was calculated as follows: Figure 13 As shown, by further extracting the spatial distribution of the internal pressure head of the 0-degree curved surface (free water surface), we can obtain... Figure 14 The results show the distribution of free water surface locations. Calculations show that the trend of groundwater flowing into the river channel from the boundary is gentler than the trend of topographic slope changes, and the free water surface on both banks is buried relatively deep. These results are consistent with the patterns observed in geological surveys. Based on the above analysis, the parameter inversion results obtained by ICSSA-RBF are acceptable. Finally, the permeability coefficients of each rock layer in the dam foundation obtained by the inversion are shown in Table 6.

[0149] Table 6

[0150]

[0151] In summary, the advantages of this invention are:

[0152] Based on Darcy's theory of unsaturated seepage, when deriving the differential equation governing unsteady unsaturated seepage, the "capacity (reflecting the dynamic relationship between pressure head and water content)" and "relative permeability (dependent on the fitting of soil-water characteristic curves)" are explicitly coupled to avoid the empirical simplification of parameters in the unsaturated zone by traditional models.

[0153] Precise description of soil-water characteristic curves: The unsaturated soil-water characteristic curves (SWCC) of the unsaturated zone are fitted using the VanGenuchten model, which has strong verification in engineering, and can accurately characterize the pore water pressure-saturation relationship in unsaturated areas such as the core wall.

[0154] The solution strategy balances accuracy and efficiency: a combined preprocessing and solution scheme of "matrix scaling + nested partitioning and sorting + LU decomposition + conjugate gradient method" is adopted, which not only solves the problem of seepage solution for large-scale 3D models, but also ensures the accuracy of calculation.

[0155] Traditional parameter inversion methods (trial and error algorithms, single intelligent algorithms) generally suffer from drawbacks such as "long iteration time and easy getting trapped in local optima," making it difficult to meet the needs of rapid acquisition of accurate hydraulic parameters in engineering. The optimized sparrow search-radial basis neural network (ICSSA-RBF) inversion framework constructed in this invention overcomes the above limitations through multi-strategy collaboration;

[0156] Logistic-Tent mapping is introduced to homogenize the initial sparrow population distribution, avoiding the "local clustering" problem of the initial samples in traditional SSA; the optimization range is expanded by combining a back-learning strategy; and Levy flight is introduced to solve the problem of individuals getting trapped in local optima during sparrow search.

[0157] Using the deviation between "head monitoring data and finite element calculation results" as the objective function, and combining actual monitoring methods such as dam body piezometers and piezometers, the permeability coefficient can be obtained by inversion, which can provide accurate input parameters for seepage simulation.

[0158] The MATLAB AppDesigner development interface allows for the rapid definition of key parameters such as permeability coefficient, porosity, and time step through intuitive controls such as "EditField input box, DropDown drop-down menu, and Slider slider".

[0159] Interactive logic adapts to engineering calculation processes: Core processes such as "assembly of saturated / unsaturated permeability matrices" and "solving by direct method / conjugate gradient method" can be triggered with a single click via a button callback function, eliminating the need for engineers to master the underlying finite element algorithms; it also supports importing boundary conditions from preprocessed files to avoid errors in complex interactive operations;

[0160] Convenient cross-environment operation and result export: The software is compiled into an independent executable file using MATLAB Compiler and packaged into an installation package using MLAppinstall, which can run stably on engineering terminals (such as computers on construction sites) without a MATLAB environment; the post-processing stage integrates statistical analysis functions, which can directly export Excel result files of hydraulic gradient and pressure head.

[0161] 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 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 claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method, characterized in that, include: S1: Develop a three-dimensional seepage calculation and analysis software platform, including deriving the unsteady unsaturated seepage control differential equation of earth-rock dam based on Darcy's unsaturated seepage theory, solving the equation based on the finite element method, and developing the software platform HYDRO-GEO / S based on the Matlab programming language. S2: Establish a three-dimensional seepage finite element model of the earth-rock dam, including establishing a three-dimensional solid model based on the earth-rock dam data and dividing it into finite element meshes, determining the initial and boundary conditions for seepage analysis, and determining the material parameters; S3: Perform three-dimensional seepage calculation and seepage safety analysis on earth-rock dams, including combining the seepage calculation program of S1 with the finite element model of S2 to perform calculations and obtain results, and evaluate the safety of the core wall, curtain grouting and other parts by comparing the hydraulic gradient calculation results with the critical hydraulic gradient. S4: Construct a parameter inversion framework to obtain the hydraulic parameters of the material, including building a Sparrow Search Algorithm-Radial Basis Neural Network (SSA-RBF) framework, optimizing it into an ICSSA-RBF framework, and combining monitoring data with finite element calculations to invert the hydraulic parameters.

2. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 1, characterized in that, S1 specifically includes: S101: Based on the principle of mass conservation, and according to the fact that the difference in inflow and outflow rates of a differential unit is equal to the rate of change in water content, the continuity equation for unsteady, unsaturated seepage can be obtained: In the formula, ρ is the density of water, and v i Where is the Darcy velocity, n is the porosity, and S is the saturation point. w For saturation, nS w =θ is the water content; S102: Unlike saturated seepage, unsaturated calculations need to consider the change in unit water content and use the pressure head h. p =p / ρg and volumetric water content θ = nS w Replace total head h and saturation S w Furthermore, the governing differential equation for unsteady, unsaturated seepage is obtained: In the formula, β is the water-holding capacity, representing the change in water content per unit volume caused by a unit change in pressure head. It is 0 in the saturated region. β is 1 in the saturated region and 0 in the unsaturated region. In hydraulic engineering, it is often taken as 0 in the saturated region as well. Then use pressure head instead. This is the ratio of relative permeability, which is 1 in the saturated region and needs to be fitted according to the soil-water characteristic curve in the unsaturated region.

3. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 2, characterized in that, S1 further includes: S102: The matrix format for deriving the global equation of seepage in the three-dimensional finite element method is as follows: In the formula, {f} is the known head at the initial boundary, [K] is the permeability matrix, [S] is the matrix reflecting the soil-water compressibility and infiltration-evaporation capacity, and [P] is the flow rate at the flow boundary. For the time term involved in the matrix format of the three-dimensional finite element global seepage equation derived from the formula, an implicit finite difference format is adopted: S103: Ignoring the effect of soil-water compression, we further obtain: S104: During the initial steady-state process calculation, it further degenerates into: [K]{h}={f} Based on the derived matrix format of the global equation for seepage in the three-dimensional finite element method, the convergence of the seepage matrix in the earth-rock dam is improved by combining matrix scaling, nested partitioning and sorting and LU decomposition preprocessing methods. Then, the conjugate gradient solution method is used to complete the iterative solution of the unsaturated seepage in the earth-rock dam.

4. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 1, characterized in that, The HYDRO-GEO / S software platform based on the Matlab programming language in S1 specifically includes: S105: In the preprocessing stage, the EditField, DropDown and Slider controls of MATLAB AppDesigner are used to define the material permeability coefficient, porosity, boundary conditions, seepage calculation type, element type and time step. A visualization window for soil-water characteristic curves is built in conjunction with the UIAxes component, and the boundary condition parameters are input through the preprocessing file. S106: In terms of interaction logic, a Button callback function is used to trigger different calculation processes, including selecting the saturated and unsaturated permeation matrix assembly mode, as well as the solution methods for large-scale linear equations using the direct method and the preprocessed conjugate gradient method. S107: In the post-processing stage, statistical analysis functions are integrated, supporting the export of calculation results of hydraulic gradient and pressure head. The App is compiled into an independent executable file using MATLAB Compiler and packaged into an installation package using MLAppinstall to ensure stable operation on engineering terminals without a MATLAB environment.

5. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 1, characterized in that, S2 specifically includes: S201: Using Ansys SpaceClaim software, based on geological exploration and design data of earth-rock dams, a three-dimensional solid model of earth-rock dams is completed through operations such as sweeping, cutting, and combining. S202: Hypermesh software was used to generate a three-dimensional finite element mesh according to the model size and accuracy requirements.

6. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 5, characterized in that, S2 further includes: The specific method for determining the initial and boundary conditions for seepage analysis of earth-rock dams is as follows: S203: Determine the first type of boundary condition, considering the unsteady seepage boundary condition related to time t, the expression is: h| Γ1 =f1(x,y,z,t) S204: Determine the second type of boundary conditions, the expression is: When considering anisotropy, the expression is: In the formula, q represents the flow rate transmitted through a unit area boundary, equivalent to v. n , l x ,l y ,l z Let n be the direction cosine between the outward normal n and the coordinate; S205: Determine the third type of boundary conditions, and establish the linear relationship between the head difference inside and outside the aquifer boundary and the exchange flow rate. The expression is as follows: In the formula, α is a positive constant, and both α and β are known values ​​at each point of this type of boundary; S206: Determine the initial conditions, and use the first type of boundary conditions as the initial state for seepage analysis.

7. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 5, characterized in that, S2 further includes: The specific method for determining the material parameters for seepage analysis of earth-rock dams is as follows: S207: Divide the finite element model materials into saturated and unsaturated regions; S208: For unsaturated materials, the relationship between pore water pressure and saturation is described using soil-water characteristic curves, fitted by the Van Genuchten model. The expression is: In the formula, Let be a function of the volumetric water content of the soil. For matrix suction, S e The degree of saturation is given by 'a', which represents the suction force corresponding to the inflection point of the soil-water characteristic curve, and 'n' represents the slope at the inflection point of the soil-water characteristic curve. 'm' = 1 - 1 / n.

8. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 1, characterized in that, S3 specifically includes: S301: Export the generated 3D finite element mesh information as a text file containing element node coordinates. The file format includes element number, node number, and node coordinates. S302: Process the mesh information text file and import the defined boundary conditions and material parameters through the software platform; S303: Perform three-dimensional seepage calculations in the HYDRO-GEO / S platform and export the hydraulic gradient and pressure head distribution results; S304: Seepage safety analysis, extract the hydraulic gradient calculation values ​​of key parts such as core wall and curtain grouting, and compare them with the critical hydraulic gradient: if the hydraulic gradient calculation value is less than or equal to the critical hydraulic gradient, the seepage prevention material is deemed safe; if the hydraulic gradient calculation value is greater than the critical hydraulic gradient, there is a risk of seepage failure.

9. The method for three-dimensional seepage simulation and parameter inversion analysis of earth-rock dams based on the finite element method according to claim 1, characterized in that, S4 specifically includes: Based on the radial basis function neural network framework and employing the sparrow search algorithm, the specific method for constructing the sparrow search algorithm-radial basis function neural network parameter inversion framework is as follows: S401: The radial basis function neural network framework includes an input layer, hidden layers, and an output layer. Based on the input vector K of the input layer, the hidden layer space constructed using radial basis functions is used to perform a nonlinear transformation on the input vector K, increasing the dimension of the input vector so that it is linearly separable in a high-dimensional space, thus linearly outputting the desired result H. This invention uses the Gaussian function, the most commonly used basis function in radial basis function neural networks, as the basis function, and its format is as follows: In the formula, K is an m-dimensional input vector; c i It is the center point of the i-th basis function with the same dimension as K; ||Kc i || is the vector norm, representing the relationship between K and c. i Euclidean distance; σ i The variance of the Gaussian function is represented by the fact that the dimension I of the center point is greater than the dimension m of the input vector, thus completing the nonlinear mapping of the input parameters from low dimension to high dimension. S402: The following formula is used to obtain the result required for the linear output of the hidden layer: In the formula, W in is the weight coefficient between output node n and the i-th center point, and N is the number of output nodes n; S403: The sparrow search algorithm is used to find the optimal parameters of the radial basis function neural network surrogate model. Based on the defined sparrow population size S, the position X(X1…X) of each sparrow is initialized. i …X q ), and calculate the fitness f(X1…X i …X q Based on the sparrow population size S, the discoverers searching for food are defined at a fixed ratio (10%–20%), and the positions of the discoverers are updated at each iteration using the following formula: In the formula, Let λ be the position of sparrow i at dimension j during the t-th iteration, and λ be a random number in the range (0,1]. max t is the maximum number of iterations; Q is a random number that follows a normal distribution; R is the warning value for the sparrow search range, which ranges from [0,1]; S T is the safe threshold for sparrows to forage, which ranges from [0.5,1]. L is a 1×q matrix with all values ​​of 1. When the warning value is less than the safe threshold, sparrows are in a state of having no natural enemies and can search for food widely. Otherwise, in order to avoid danger, sparrows will move randomly in a normal distribution and forage in a small area. S404: In a sparrow population, the followers other than the discoverer constitute the main group. They obtain food based on the discoverer's location, and there is a mutual surveillance relationship among the followers. Followers will always tend to approach the location with the highest food availability, and their locations are updated according to the following strategy: In the formula, It is the most unfavorable position in the t-th iteration; It is the optimal position in the (t+1)th iteration; A + =A T (AA T ) -1 It is the Moore-Penrose generalized inverse of matrix A, where A is a matrix consisting of -1 and 1. When i > n / 2, it means that the follower has difficulty obtaining food and needs to search for other locations to forage; otherwise, it means that the follower will compete with the discoverer for food and may replace the discoverer's position. S405: Within a sparrow population, 10%–20% are vigilant individuals. When they detect danger, they will promptly adjust their positions to avoid it. The vigilant individual position update strategy is as follows: In the formula, It is the optimal position in the t-th iteration; f i This is the current fitness value of the sparrow; f g It is the fitness value of the current global best position; f w is the fitness value of the current globally unfavorable position; β is a random number that follows a standard normal distribution; η is a random number in the range [0,1]; ε is a constant to avoid a denominator of 0, f i >f g This means that sparrows are on the edge of the population and are more vulnerable to predators; f i =f g When this happens, it indicates that the sparrows in the middle of the population are aware of the threat from predators and will move to avoid the risk. S406: The specific method for constructing the optimized sparrow search-radial basis neural network inversion framework by introducing a reverse learning strategy, Lévy flight random walk, Logistic-Tent mapping, and variable step size search optimization strategy is as follows: S407: Introducing a reverse learning strategy, based on the current optimal solution, obtaining the corresponding reverse solution, comparing and evaluating the reverse solution and the optimal solution, and updating the current optimal solution, in order to expand the optimization ability and scope of each individual in the population. The strategy is as follows: In the formula, ub and lb are the proposed upper and lower boundary values, respectively; yes The inverse solution; r is a 1×q random matrix uniformly distributed in the range (0,1); b=(Tt / T) t These are control parameters related to information exchange; S408: Lévy flight is introduced to solve the problem of sparrows getting trapped in local optima during search. The sparrow random walk strategy after Lévy flight perturbation is as follows: S409: To address the search objectives before and after the initial search, a position update weight is introduced, and the discoverer's position update strategy is updated as follows: In the formula, ω t ω0 is the weight of the current iteration step t; ω0 is a constant greater than 1; t0 is the number of iterations at which the local search capability is enhanced. S4010: The initial sparrow samples are optimized using Logistic-Tent mapping to make their distribution more uniform. The mapping relationship of the initial samples is as follows: In the formula, r is a random number in (0,4). Based on the updated sparrow search method described above, the search for the optimal radial basis parameters is completed quickly and efficiently, and finally the optimized sparrow search-radial basis neural network inversion framework is constructed. The specific method for obtaining the hydraulic parameters of materials by combining monitoring data and finite element calculation inversion is as follows: S4011: Determine the evaluation indicators for inversion analysis. Obtain monitoring data for the evaluation indicators through exploration. The water head is the most easily measurable hydraulic parameter of the natural foundation and dam structure at the dam site, and it is also a hydraulic indicator that can be directly calculated by seepage finite element analysis. Therefore, establish the objective function of the radial basis function based on the water head: In the formula, H n It is the calculated pressure head value at the target location n; H n ′ represents the inversion head value at the target location n; This is the weight at the target position n, reflecting the importance of the data at that point within the overall data. When the importance of nodes is the same, the weight is taken as the highest value. S4012: Based on the finite element model established in S4-S6, the developed seepage calculation and analysis platform is called within the ICSSA-RBF framework to perform finite seepage analysis and complete the inversion of hydraulic parameters to obtain the required hydraulic parameters.

10. A three-dimensional seepage simulation and parameter inversion analysis system for earth-rock dams based on the finite element method, applicable to the method described in any one of claims 1-9, characterized in that, include: Software platform development module, finite element model construction module, seepage calculation and safety analysis module, parameter inversion module; The software platform development module is used to develop a three-dimensional seepage calculation and analysis software platform HYDRO-GEO / S with pre-processing, calculation and post-processing functions based on Darcy's unsaturated seepage theory, finite element method and Matlab language. The finite element model building module is used to complete three-dimensional solid modeling and mesh generation based on earth-rock dam data, define the initial conditions, boundary conditions and material parameters for seepage analysis, and include soil-water characteristic curves fitted by the VanGenuchten model. The seepage calculation and safety analysis module is used to import the finite element model into the HYDRO-GEO / S platform to perform three-dimensional seepage calculation, export the hydraulic gradient results and compare them with the critical hydraulic gradient to evaluate the seepage safety of the core wall, curtain grouting, etc. The parameter inversion module is used to build an SSA-RBF framework and optimize it into an ICSSA-RBF framework. Combined with head monitoring data and finite element calculations, it inverts and obtains the hydraulic parameters of earth-rock dam materials.

Images