Long-term safety evaluation method for high dam engineering based on data and mechanism mixed driving

Abstract

This invention discloses a long-term safety evaluation method for high dam projects based on a data- and mechanism-driven approach, relating to the fields of hydraulic engineering and geotechnical mechanics. The method includes: constructing an integrated high dam-dam foundation model using professional modeling software based on geological survey data and high dam construction data, and conducting finite element analysis; simultaneously acquiring model constraints, initial mechanical parameters, and on-site displacement and stress monitoring data of the dam body; inputting the above conditions and parameters into the model for trial calculation to verify their rationality; simultaneously, relying on the monitoring data, obtaining optimized mechanical parameters through machine learning algorithms; after the model verification is successful, calculating the dam body displacement and stress using the optimized parameters; if unsuccessful, returning to remodel and recalculate; comparing the error between the model calculation values ​​and the on-site monitoring values; if the error meets the threshold, performing displacement and stress prediction; otherwise, recalculating, ultimately achieving an accurate evaluation of the long-term safety of the high dam project. This invention improves the engineering safety of the dam body.

Description

Technical Field

[0001] This invention relates to the fields of hydraulic engineering and geotechnical mechanics, and in particular to a method and apparatus for long-term safety evaluation of high dam projects based on a hybrid data and mechanism approach. Background Technology

[0002] my country's hydropower resources are distributed unevenly across regions. With its abundant hydropower reserves and unique mountain and canyon topography, the southwest region has become the core area for hydropower project construction. The development of its hydropower resources is of great significance for optimizing the national energy structure and alleviating the contradiction between energy supply and demand.

[0003] High dam projects have service lives spanning decades or even centuries, and the geomechanical properties of the dam foundation and reservoir soil are not constant. During long-term operation, the coupled effects of multiple factors continuously alter regional geological conditions: long-term seepage easily leads to rock softening and increased weathering depth; periodic fluctuations in reservoir water levels constantly disturb the seepage and stress fields of the bank slopes and dam foundation. These factors collectively cause dynamic changes in key parameters of the strata beneath the dam foundation, including mechanical indicators such as rock deformation modulus and shear strength, the spatial location and infill properties of faults and weak interlayers, and the development depth of weathering unloading zones, directly threatening the long-term stability and safety of the dam.

[0004] Traditional engineering geological survey and geotechnical mechanics testing methods have revealed significant limitations after the project is completed and water is impounded. During the dam operation phase, it is difficult to obtain representative soil and rock samples reflecting the current state and conduct mechanical tests through conventional methods such as supplementary drilling, on-site sampling, and in-situ testing; relying solely on preliminary survey data is no longer sufficient to accurately characterize the dam's true working condition.

[0005] Against this backdrop, conducting mechanical parameter inversion based on existing monitoring data and combining it with numerical simulation technology to predict key indicators such as dam displacement and stress has become an important technical approach for long-term safety evaluation of high dam projects, and is of great practical significance for ensuring the safe service of high dam projects. Summary of the Invention

[0006] The main objective of this invention is to provide a long-term safety evaluation method for high dam projects based on a hybrid data and mechanism approach.

[0007] Another objective of this invention is to propose a long-term safety evaluation device for high dam projects based on a hybrid data and mechanism approach.

[0008] The third objective of this invention is to provide an electronic device.

[0009] A fourth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0010] To achieve the above objectives, a first aspect of the present invention proposes a long-term safety evaluation method for high dam projects based on a data- and mechanism-driven approach, comprising:

[0011] S1. Based on geological survey data and high dam construction data, an integrated model of high dam and dam foundation is constructed. The constraints and initial mechanical parameters of the integrated model of high dam and dam foundation are obtained through finite element analysis. On-site monitoring of the dam body is carried out, and dam body displacement monitoring data and dam body stress monitoring data are collected. S2, Input the model constraints and initial mechanical parameters into the integrated high dam-dam foundation model for trial calculation to verify the rationality of the model. At the same time, based on the dam body displacement and stress monitoring data, use machine learning algorithms to perform parameter inversion and obtain optimized mechanical parameters. S3. If the integrated high dam-dam foundation model verification is qualified, the optimized mechanical parameters will be input into the model to calculate the dam displacement and stress. If the model verification is unqualified, the model construction and trial calculation verification process will be repeated. S4 compares and analyzes the displacement and stress data of the dam body calculated by the model with the data monitored on site, obtains and analyzes the error. If the error meets the preset threshold, the displacement and stress of the dam body are predicted based on the optimized mechanical parameters. Otherwise, the parameter inversion process is repeated to realize the evaluation of the long-term safety of the high dam project.

[0012] Optionally, the construction of the integrated high dam-dam foundation model further includes: Based on geological survey data, contour maps of the dam foundation area were obtained to determine the modeling scope and to analyze the stratigraphic structure, joint and fracture development characteristics. Modeling software was used to create the original mountain model, joint and fissure model, and dam structure model, respectively. Each sub-model is divided into grids and integrated to form an integrated high dam-dam foundation model.

[0013] Optionally, the step of obtaining the constraint conditions and initial mechanical parameters of the integrated high dam-dam foundation model through finite element analysis further includes: Determine the initial conditions, boundary conditions, and external loads of the model, wherein the external loads include water pressure and sediment pressure; Physical and soil mechanics tests were conducted to obtain the initial mechanical parameters required for each model. These initial mechanical parameters included density, elastic modulus, Poisson's ratio, tensile strength, cohesion, and internal friction angle.

[0014] Optional, verifying the model's rationality includes: The obtained initial conditions, boundary conditions, external loads, and initial mechanical parameters of the model are input into the integrated high dam-dam foundation model for trial calculation; The verification is determined by judging whether the model converges during the trial calculation. If the model does not converge, the verification is deemed unqualified; if the model converges, the verification is deemed qualified.

[0015] Optionally, the step of using machine learning algorithms to perform parameter inversion and obtain optimized mechanical parameters includes: The dam displacement and stress monitoring data are preprocessed, and the preprocessed monitoring data are used to perform parameter inversion using a preset machine learning inversion algorithm to obtain the optimized mechanical parameters required for each model. The range of values ​​for the optimized mechanical parameters is defined by the initial mechanical parameters obtained from the experiment, ensuring the rationality of the optimized mechanical parameters.

[0016] Optionally, the process for determining whether the error meets a preset threshold includes: The error judgment threshold is set at 10% of the monitoring data. The error results of the dam displacement and stress data calculated by the model and the field monitoring data are compared with this threshold. If the error result is greater than the threshold, return to the process corresponding to parameter inversion, adjust the inversion model parameters and optimize the model before re-performing parameter inversion; If the error result is less than the threshold, the subsequent displacement and stress prediction process begins. To achieve the above objective, a second aspect of this invention proposes a long-term safety evaluation device for high dam projects based on a data- and mechanism-driven hybrid approach, comprising: The modeling and monitoring module is used to construct an integrated model of the high dam and its foundation based on geological survey data and high dam construction data. It obtains the constraints and initial mechanical parameters of the integrated model of the high dam and its foundation through finite element analysis, and conducts on-site monitoring of the dam body, collecting dam body displacement monitoring data and dam body stress monitoring data. The trial calculation and inversion module is used to input the model constraints and initial mechanical parameters into the integrated high dam-dam foundation model for trial calculation to verify the rationality of the model. At the same time, based on the dam body displacement and stress monitoring data, machine learning algorithms are used to carry out parameter inversion to obtain optimized mechanical parameters. The verification calculation module is used to input optimized mechanical parameters into the model and calculate the dam displacement and stress if the high dam-dam foundation integrated model verification is qualified; if the model verification is unqualified, the model construction and trial calculation verification process is repeated. The error evaluation module is used to obtain and analyze the dam displacement and stress data calculated by the comparative analysis model and the field monitoring data. If the error meets the preset threshold, the dam displacement and stress are predicted based on the optimized mechanical parameters. Otherwise, the parameter inversion process is repeated to realize the evaluation of the long-term safety of the high dam project.

[0017] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0018] To achieve the above objectives, a third aspect of this application provides an electronic device, including a processor and a memory; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, for implementing a long-term safety evaluation method for high dam projects based on a data and mechanism hybrid drive as described in the first aspect embodiment.

[0019] To achieve the above objectives, the fourth aspect of this application proposes a non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, implements a long-term safety evaluation method for high dam projects based on a data- and mechanism-driven hybrid approach as described in the first aspect embodiment.

[0020] The embodiments of this invention have the following beneficial effects: This invention provides a long-term safety evaluation method for high dam projects based on a data- and mechanism-driven approach. It integrates physical mechanism models with on-site monitoring data, effectively solving the key problem of directly obtaining changes in the mechanical parameters of soil and rock masses during long-term dam operation. Simultaneously, this method uses machine learning algorithms to integrate monitoring data for parameter calibration, inverting and obtaining mechanical parameters that better reflect the actual state of the project. Numerical simulation software is then used for verification and prediction, forming a complete process with minimal error. This not only significantly improves the accuracy and reliability of the safety evaluation model but also enables scientific prediction and risk warning of future stress and displacement development in the dam body. It provides efficient and scientific decision support for the long-term safe operation and maintenance and risk prevention of high dam projects, possessing significant engineering application value. Attached Figure Description

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart of a long-term safety evaluation method for high dam projects based on a hybrid data and mechanism approach provided in this embodiment of the invention; Figure 2 A flowchart illustrating the overall framework of a long-term safety evaluation method for high dam projects based on a hybrid data and mechanism approach, provided in this embodiment of the invention. Figure 3 The two-dimensional mesh model provided in the embodiments of the present invention; Figure 4 This is a schematic diagram of monitoring data provided in an embodiment of the present invention; Figure 5 The displacement-stress cloud diagram of the two-dimensional model provided in the embodiments of the present invention; Figure 6 The principal stress cloud diagram of the two-dimensional model provided in the embodiment of the present invention; Figure 7 The small principal stress cloud diagram of the two-dimensional model provided in the embodiment of the present invention; Figure 8 Damage stress cloud diagram provided for embodiments of the present invention; Figure 9 This is a structural diagram of a long-term safety evaluation device for high dam projects based on a hybrid data and mechanism approach, provided in an embodiment of the present invention. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] To enable those skilled in the art to better understand the present invention, 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] The following description, with reference to the accompanying drawings, describes a method and apparatus for long-term safety evaluation of high dam projects based on a hybrid data and mechanism approach, according to an embodiment of the present invention.

[0025] Example 1 This invention provides a long-term safety evaluation method for high dam projects based on a hybrid data and mechanism approach. Figure 1 This is a flowchart illustrating a long-term safety evaluation method for high dam projects based on a data- and mechanism-driven approach, as provided in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the overall framework of a data- and mechanism-driven long-term safety evaluation method for high dam projects, as provided in this embodiment of the invention. Figure 1 , Figure 2 As shown, the method includes the following steps: Step S1: Based on geological survey data and high dam construction data, construct an integrated high dam-dam foundation model, obtain the constraint conditions and initial mechanical parameters of the integrated high dam-dam foundation model through finite element analysis, and conduct on-site monitoring of the dam body to collect dam body displacement monitoring data and dam body stress monitoring data.

[0026] First, this embodiment of the application uses geological survey data and high dam construction data as its core foundation to carry out preliminary data sorting and scope definition work. Specifically, this embodiment of the application conducts systematic geological surveys of the dam foundation area, and combines existing survey reports, exploration records, and other data to accurately draw contour maps at different elevations of the dam foundation, clearly presenting the topographic features of the dam foundation area. At the same time, this embodiment of the application strictly reviews industry specifications, technical standards, and relevant domestic and foreign research literature related to high dam engineering, and reasonably determines the overall scope of modeling based on the actual needs of the project. In the completed contour map, a specific scope suitable for integrated high dam-dam foundation modeling is accurately selected to avoid the problems of computational redundancy caused by an excessively large modeling scope and model distortion caused by an excessively small modeling scope. In addition, this embodiment of the application also comprehensively summarizes the stratigraphic structure and joint development within the selected scope, clarifying the stratigraphic distribution pattern, lithological variation characteristics, and key information such as joint development density and distribution direction, providing comprehensive geological foundation support for subsequent model construction.

[0027] Secondly, based on the geological and engineering data compiled above, this embodiment of the application utilizes professional modeling software to construct an integrated high dam-dam foundation model, and completes mesh generation and import into numerical simulation software. The specific process is as follows: First, based on core geological parameters such as the initial lithology, weathering degree, and thickness of each stratum at the dam foundation location, an original mountain model is accurately established to realistically reproduce the geological structural characteristics of the mountain where the dam foundation is located. Subsequently, for geological defects such as faults, joints, and fissures in the mountain, key parameters such as their strike, thickness, dip angle, dip direction, and depth range are accurately obtained through geological survey data, and joint and fissure models are established one by one. To ensure the model accurately reflects the impact of geological defects in the mountain on the stress of the dam body, this embodiment of the application, based on construction data such as high dam construction design drawings and construction records, establishes sub-models of the dam body according to construction segments or structural segments. Then, using professional modeling techniques, these sub-models are combined and spliced ​​together to form a complete dam body model, ensuring that the dam body model is completely consistent with the actual engineering structure. Finally, this embodiment of the application performs unified mesh generation on the established original mountain model, joint and fissure model, and dam body model, combining each model into a complete integrated high dam-dam foundation model. The mesh file is exported and imported into numerical simulation software, laying the foundation for subsequent finite element analysis.

[0028] In the mesh generation process, this embodiment of the application can achieve gradual mesh generation according to the computational accuracy requirements of different regions, balancing computational efficiency and accuracy: for areas with concentrated stress, such as the dam body and faults, where computational accuracy is highly demanding, a smaller mesh size and higher density are used to ensure accurate capture of stress and displacement characteristics in these areas; for areas in the mountain model far from the dam body and less affected by stress, the mesh size can be appropriately increased, the mesh density reduced, and the computational load decreased, achieving a balance between computational efficiency and accuracy. Furthermore, the numerical simulation software used in this embodiment of the application, including but not limited to three-dimensional numerical calculation programs suitable for geotechnical engineering and structural analysis, can be flexibly selected according to actual engineering needs. The modeling and calculation process includes both two-dimensional and three-dimensional modeling and calculation scenarios, adapting to the evaluation needs of high dam projects of different scales and with varying accuracy requirements.

[0029] Furthermore, this embodiment uses finite element analysis to obtain the constraint conditions and initial mechanical parameters A of the integrated high dam-dam foundation model. Specifically, this embodiment uses finite element analysis to clarify the initial conditions (such as initial stress state, initial displacement state, etc.), boundary conditions (such as fixed boundaries, free boundaries, etc.), and external load conditions of the model. The external loads mainly include common load types encountered during the operation of high dam projects, such as water pressure and sediment pressure, comprehensively covering all factors influencing the model's stress. The finite element static equilibrium governing equations are:

[0030] Where [K] is the overall stiffness matrix of the integrated high dam-dam foundation model; {u} is the displacement vector of the model nodes; and {F} is the external load vector, including self-weight, water pressure, sediment pressure, etc.

[0031] The initial geostress field satisfies the self-weight stress distribution:

[0032] in, γ represents the initial vertical ground stress at the calculation point; γ represents the natural unit weight of the soil and rock mass; and z represents the depth of the calculation point from the ground surface.

[0033] Meanwhile, this application embodiment conducts indoor physical tests, soil mechanics tests and other related tests to accurately determine the mechanical parameters of various rock and soil materials and structural materials of the dam body and dam foundation, including density, elastic modulus, Poisson's ratio, tensile strength, cohesion and internal friction angle, etc. The parameters obtained from these tests are used as initial mechanical parameters A to provide initial benchmarks for model calculation and parameter inversion.

[0034] At the same time, the embodiments of this application simultaneously carry out on-site monitoring of the dam body. By deploying monitoring equipment at key parts of the dam body, real-time monitoring data on dam body displacement and stress are collected. The monitoring range covers key sections such as the top, middle and bottom of the dam body, as well as areas of concentrated stress, ensuring that the monitoring data can comprehensively and accurately reflect the actual working state of the dam body, and providing reliable on-site data support for subsequent model verification, parameter inversion and error analysis.

[0035] Step S2: Input the model constraints and initial mechanical parameters into the integrated high dam-dam foundation model for trial calculation to verify the model's rationality. Simultaneously, based on the dam body displacement and stress monitoring data, use machine learning algorithms to perform parameter inversion and obtain optimized mechanical parameters.

[0036] First, this embodiment of the application conducts model trial calculations and rationality verification to ensure that the constructed integrated high dam-dam foundation model has engineering application value. Specifically, this embodiment of the application inputs all the model constraints (including initial conditions, boundary conditions, external loads, etc.) obtained in step S1 and the initial mechanical parameters A obtained from the experiment into the constructed integrated high dam-dam foundation model and starts the model trial calculation process. The core purpose of the model trial calculation is to verify the rationality of the model. The specific judgment criterion is whether the model calculation process converges. If the model calculation converges successfully, it means that the model's structural construction, mesh generation, parameter settings, etc., all conform to engineering reality and mechanical laws, and the model verification is qualified, and it can proceed to the subsequent parameter inversion stage. If the model calculation fails to converge, it indicates that there are unreasonable aspects in the model. At this time, this embodiment of the application needs to return to the model construction stage, adjust the modeling order, optimize the mesh distribution (such as adjusting the mesh density and correcting the mesh gradient method), and restart the model construction and trial calculation work until the model can converge successfully and pass the verification, ensuring the accuracy of subsequent parameter inversion and numerical calculation.

[0037] Secondly, this embodiment preprocesses the dam displacement and stress monitoring data collected in step S1 to remove data interference and improve data quality for subsequent parameter inversion. In this embodiment, monitoring data preprocessing is an important guarantee for the accuracy of parameter inversion. The specific operation is as follows: This embodiment first classifies and organizes the collected monitoring data according to monitoring time and monitoring location (such as the top, middle, bottom, and key stress sections of the dam), clarifying the monitoring data characteristics of different time periods and locations; then, outliers and missing values ​​are identified and processed in the classified monitoring data. Abnormal fluctuation values ​​and invalid data in the monitoring data are identified and removed using professional data processing methods. At the same time, reasonable interpolation methods are used to supplement missing values ​​to avoid interference from missing values ​​and outliers in the parameter inversion results.

[0038] Missing monitoring data were completed using linear interpolation. The interpolation formula is as follows:

[0039] in, , Given the known monitoring time, , For the corresponding monitoring value, x is the time to be interpolated, and y is the interpolation result.

[0040] Furthermore, the embodiments of this application also construct characteristic parameters with clear physical meaning from the original monitoring data, such as dam deformation rate, displacement time-dependent components, and water level change gradient. These constructed features can more comprehensively reflect the stress and deformation law of the dam body, providing richer and more effective data support for subsequent parameter inversion, and further improving the reliability of the inversion results.

[0041] Meanwhile, this application embodiment establishes a mechanistic model to constrain the parameter inversion process, avoiding the potential failure of physical meaning in purely data-driven inversion, and ensuring that the inversion parameters conform to the actual mechanical laws of engineering. In this application embodiment, to meet the calculation requirements of the dam body's rheological properties, a Burgers rheological model is established. This model can accurately describe the viscoelastic rheological behavior of the dam body's soil and rock, and its expression is:

[0042] in, The total strain of the soil and rock mass at time t; It is the instantaneous elastic modulus, which mainly reflects the elastic deformation capacity of rock and soil under instantaneous load; The viscosity coefficient characterizes the viscous deformation properties of soil and rock masses. The delayed elastic modulus reflects the delayed elastic deformation characteristics of soil and rock under long-term loading. The delayed viscosity coefficient describes the development law of delayed viscous deformation in soil and rock masses; The constant stress level is denoted by t; time is t. This embodiment of the application, through the mechanistic constraints of the Burgers rheological model, integrates the rheological laws of soil and rock into the parameter inversion process. This ensures that parameter inversion no longer relies solely on the statistical laws of monitoring data, but rather combines engineering mechanics mechanisms. This effectively avoids the problems that may occur in pure data-driven inversion, such as inversion parameters not matching actual mechanical laws and the loss of physical meaning, thus providing scientific mechanistic support for parameter inversion.

[0043] In the third stage, this application embodiment conducts machine learning-driven parameter inversion work to obtain optimized mechanical parameters B by combining mechanistic constraints. Specifically, technicians first preset suitable machine learning inversion algorithms, such as support vector machines and neural networks, based on the actual needs of the high dam project and the characteristics of the monitoring data, to ensure that the algorithms can accurately capture the complex mapping relationship between monitoring data and mechanical parameters. Subsequently, the preprocessed monitoring data (including the original monitoring values ​​and the characteristic parameters of the structure) is input into the preset machine learning inversion algorithm. Through multiple rounds of iterative learning of the algorithm, an accurate mapping relationship between the monitoring data and the mechanical parameters required by each model of the dam body and dam foundation is gradually established, thereby inverting to obtain the mechanical parameters B corresponding to each module.

[0044] Meanwhile, to further standardize the parameter inversion process and prevent the inverted mechanical parameter B from deviating from common sense in engineering mechanics, this application's embodiment introduces the initial mechanical parameter A obtained through physical and soil mechanics tests in step S1 as a constraint condition, clarifying the range of values ​​for mechanical parameter B. For example, the inverted values ​​of key mechanical parameters such as elastic modulus, cohesion, and internal friction angle must not exceed the reasonable range corresponding to the initial mechanical parameter A. Through this combination of mechanism constraints and data-driven approach, the parameter inversion process is further standardized, ensuring that the inverted mechanical parameter B not only conforms to the actual state reflected by the monitoring data but also adheres to the laws of engineering mechanics. This significantly improves the reliability and accuracy of parameter inversion, providing core parameter support for the subsequent numerical calculations in step S3 and the safety evaluation in step S4, ensuring the scientific validity and practicality of the long-term safety evaluation method for the entire high dam project.

[0045] Step S3: If the integrated high dam-dam foundation model verification is qualified, the optimized mechanical parameters are input into the model to calculate the dam displacement and stress. If the model verification is unqualified, the model construction and trial calculation verification process is repeated.

[0046] In the embodiments of this application, S3 is a crucial step in the long-term safety evaluation method for high dam projects, which connects the preceding and following steps. Its core function is to clarify the implementation path of subsequent numerical calculations based on the model verification results of step S2, ensure the rationality of model calculations and the accuracy of data, and lay a solid foundation for error analysis and safety evaluation in step S4.

[0047] First, this application embodiment clarifies the core judgment criterion for step S3—namely, the verification result of the integrated high dam-dam foundation model in step S2. This verification result directly determines the execution direction of step S3 and is an important prerequisite for ensuring the rigor of the entire evaluation process. In this application embodiment, the core standard for model verification is the convergence of the model trial calculation in step S2. If the model trial calculation converges successfully and conforms to the actual mechanical laws of engineering after mesh adjustment and parameter calibration, the integrated high dam-dam foundation model is deemed to be qualified and ready for subsequent numerical calculations. If the model trial calculation fails to converge, or still does not meet the actual engineering requirements after multiple adjustments, the model verification is deemed unqualified, and the model construction and trial calculation verification work needs to be carried out again.

[0048] Secondly, for cases where the model verification is successful, this embodiment of the application executes the numerical calculation process. Specifically, after the integrated high dam-dam foundation model is verified as successful, this embodiment of the application will input the optimized mechanical parameters B obtained by machine learning algorithm inversion in step S2 into the successful model and start the model numerical calculation process. The core objective of this numerical calculation is to accurately obtain the displacement and stress distribution data of the dam body under actual operating conditions. The calculation process strictly follows the finite element analysis principle, and combines the model constraints, external loads and other parameters determined in step S1 to comprehensively simulate the stress state of the dam body and dam foundation. The calculation focuses on the displacement and stress values ​​of key sections of the dam body and stress concentration areas (such as the bottom of the dam body and near faults) to ensure that the calculated data can truly reflect the actual stress and deformation characteristics of the dam body, providing reliable model calculation data support for subsequent comparative analysis with on-site monitoring data.

[0049] Finally, in cases where model verification fails, this embodiment of the application initiates a model reconstruction and re-verification process. In this embodiment, if the integrated high dam-dam foundation model verification fails, it indicates that there are still unreasonable aspects in the model's structural construction, mesh generation, parameter settings, etc. Directly proceeding to numerical calculations would lead to distorted calculation results, failing to provide effective support for safety evaluation. Therefore, this embodiment of the application needs to return to the model trial calculation and verification stage in step S2, and re-construct the model, including adjusting the modeling order, optimizing the mesh distribution (such as correcting mesh density and adjusting mesh gradation methods), calibrating model constraints, etc. After completing model reconstruction, the trial calculation and verification process is executed again until the model trial calculation converges and verification is successful before proceeding to the subsequent numerical calculation stage. This ensures the scientific rigor and soundness of the entire evaluation process, avoiding deviations in subsequent evaluation results due to unreasonable models.

[0050] Step S4 involves comparing and analyzing the displacement and stress data of the dam body calculated by the model with those monitored on-site to obtain and analyze the error. If the error meets the preset threshold, the displacement and stress of the dam body are predicted based on the optimized mechanical parameters. Otherwise, the parameter inversion process is repeated to evaluate the long-term safety of the high dam project.

[0051] First, this application embodiment conducts a comparative analysis of model calculation data and field monitoring data to accurately obtain the error δ and complete error analysis. This is a crucial prerequisite for judging the effectiveness of parameter inversion and ensuring the reliability of subsequent prediction results. In this application embodiment, the core objects of the comparative analysis are the dam displacement and stress data obtained through numerical calculation in step S3, and the dam displacement and stress data obtained through field monitoring in step S1. The comparison scope comprehensively covers the key sections of the dam, the stress concentration areas, and all monitoring points to ensure the comprehensiveness and representativeness of the comparison results. Specifically, the relative error calculation formula is used:

[0052] in, This represents the relative error between the model-calculated values ​​and the measured values. The displacement or stress values ​​obtained from the finite element model; These are the measured displacement or stress values ​​obtained from on-site monitoring.

[0053] Specifically, this application embodiment employs a scientific error calculation method, comparing model calculation data and on-site monitoring data at the same monitoring point and within the same time period to calculate the error δ corresponding to each point and parameter. Simultaneously, it comprehensively analyzes the magnitude, distribution pattern, and causes of the error δ, clarifying whether the error originates from parameter inversion deviation, unreasonable model settings, or interference from monitoring data, providing a clear basis for subsequent process adjustments. This application embodiment presets an error judgment threshold. Combining industry standards for high dam engineering safety evaluation with actual engineering needs, the error threshold is set at 10% of the monitoring data. That is, when the error δ > 10%, it indicates that the optimized mechanical parameter B obtained in step S2 still has deviations, and the parameter inversion process has not met the engineering accuracy requirements. In this case, it is necessary to return to step S2, adjust the machine learning inversion algorithm parameters, optimize the data preprocessing process, and restart the parameter inversion work until the inverted mechanical parameter B can make the error meet the preset threshold. When the error δ ≤ 10%, it indicates that the parameter inversion result is accurate and the model calculation is reliable, and the subsequent displacement and stress prediction stages can proceed.

[0054] Secondly, under the premise that the error δ meets the preset threshold (δ≤10%), this embodiment of the application conducts long-term displacement and stress prediction of the dam body based on the optimized mechanical parameter B, providing a core basis for the long-term safety evaluation of high dam projects. Specifically, this embodiment of the application inputs the optimized mechanical parameter B obtained from the inversion in step S2 back into the verified high dam-dam foundation integrated model, and uses numerical analysis software to start the long-term numerical calculation process to comprehensively simulate the stress and deformation process of the high dam under long-term operating conditions (such as long-term water pressure and sediment pressure), and accurately calculate the displacement and stress distribution of the dam body at different time points and under different operating conditions. The long-term time-dependent displacement of the dam body can be expressed as:

[0055] Where u(t) is the total time-dependent displacement of the dam body at time t; It is an instantaneous elastic displacement, generated the instant the load is applied; To delay elastic displacement, it gradually tends to stabilize over time; It is a viscous rheological displacement that develops approximately linearly with time.

[0056] After calculation, this embodiment of the application derives a cloud map of dam displacement and stress distribution. The cloud map clearly presents the displacement and stress distribution characteristics of various regions of the dam, highlighting key areas with excessive displacement and stress concentration. This provides intuitive and accurate data support for subsequent long-term safety analysis, hazard identification, and operation and maintenance decisions for high dam projects. Through this prediction process, the stress and deformation trends of high dams during long-term operation can be predicted in advance, potential safety hazards can be identified in a timely manner, and a scientific and accurate evaluation of the long-term safety of high dam projects can be achieved. This fully demonstrates the advantages of the data-mechanism hybrid driving method of this application, effectively solving the problems of insufficient accuracy and poor alignment with engineering realities in traditional evaluation methods.

[0057] Finally, through the above-mentioned error analysis, parameter calibration, and long-term prediction process, the embodiments of this application form a complete closed loop for the long-term safety evaluation of high dam projects, ensuring the scientific, accurate, and practical nature of the evaluation results. This provides reliable technical support for the safe operation and maintenance, hidden danger management, and long-term stable operation of high dam projects, effectively guaranteeing the operational safety of high dam projects.

[0058] In the application of one embodiment of the present invention, the implementation process is as follows: This implementation case will be demonstrated using a two-dimensional mesh model; like Figure 3 As shown, the modeling range required based on geological survey data and high dam construction data integration is: x range 350m, y range 237m, thickness 10m to establish a two-dimensional mesh model. This model is discretized using triangular elements, which can clearly reflect the geometric shape and material zoning characteristics of the dam body and surrounding foundation. Table 1

[0059] Table 1 shows the mechanical parameters A for different materials. Figure 4 The figure clearly shows the displacement time-history curves of monitoring points 1, 2, and 3 over 0 to 360 days, based on displacement data from different monitoring points. All three curves exhibit a trend of slow increase over time followed by gradual stabilization, with monitoring point 2 showing the largest displacement and monitoring point 1 showing the smallest. The displacement and stress data obtained from monitoring the dam were processed and analyzed, outliers were removed, and a Burgers-Mohr rheological model was established using machine learning. The mechanical parameter B, calibrated by fusion data, is shown in Table 2.

[0060] Table 2

[0061] Input the inverted mechanical parameter B into the model, and perform seepage stress coupled elastoplastic calculation and seepage stress coupled rheological calculation on the model to analyze its dam displacement, major and minor principal stress data, etc.

[0062] like Figure 5 The figure shown is a displacement-stress cloud diagram of the two-dimensional model of this invention. The cloud diagram visually displays the displacement distribution of the dam body and foundation, with a maximum displacement value of approximately 2.19 × 10⁻⁶. - The data is mainly concentrated on the upper part of the dam body and the abutment. The specific locations of monitoring point 1, monitoring point 2 and monitoring point 3 are marked on the map to facilitate comparison with the measured data.

[0063] like Figure 6 The figure shown is a contour map of the major principal stresses in the two-dimensional model of this invention. The contour map reflects the distribution of major principal stresses in the dam body and foundation, with stress values ​​ranging from -4.74 × 10⁻⁶. 4 Pa to -8.00×10 6 Pa, stress concentration is more pronounced at the contact points between the dam body and the foundation, as well as at the dam shoulders.

[0064] like Figure 7 The figure shown is a contour plot of the minor principal stresses in the two-dimensional model of this invention. The contour plot illustrates the distribution characteristics of the minor principal stresses, with stress values ​​ranging from 1.20 × 10⁻⁶. 5 Pa to -2.50×10 6 Pa indicates a relatively large gradient in the minor principal stress distribution on the downstream side of the dam and in the shallow foundation region. The calculated displacement data is compared and analyzed with the given data.

[0065] in, This represents the relative error between the model-calculated values ​​and the measured values. The displacement or stress values ​​obtained from the finite element model; These are the measured displacement or stress values ​​obtained from on-site monitoring.

[0066] Monitoring point 1: δ1 = (7.0 - 6.5) / 6.55 = 0.078 ≤ 10%; Monitoring point 2: δ2 = (24-22) / 24 = 0.0833 ≤ 10%; Monitoring point 3: δ3 = (17.5 - 17) / 17 = 0.029 ≤ 10%; This indicates that the parameter inversion was successful and can provide mechanical parameters for subsequent predictions.

[0067] Finally, the model is subjected to seepage stress coupling elastoplastic calculations, such as... Figure 8 The image shown is a stress cloud diagram of the dam body damage according to the present invention. The cloud diagram clearly shows the damage distribution of the dam body, with damage values ​​ranging from 0 to 9.86 × 10⁻⁶. - ² The damage to the bottom of the dam body and the local area in contact with the foundation is relatively high. Based on this, a dam damage model can be predicted, providing a more scientific theoretical basis for subsequent dam stability analysis and dam reinforcement.

[0068] Example 2 This invention provides a long-term safety evaluation device for high dam projects based on a hybrid data and mechanism approach. Figure 9 This is a schematic diagram of a long-term safety evaluation device for high dam projects based on a hybrid data and mechanism approach, provided in an embodiment of the present invention. Figure 9 As shown, the device includes: The modeling and monitoring module 100 is used to construct an integrated model of the high dam and its foundation based on geological survey data and high dam construction data. It obtains the constraints and initial mechanical parameters of the integrated model of the high dam and its foundation through finite element analysis, and conducts on-site monitoring of the dam body, collecting dam body displacement monitoring data and dam body stress monitoring data. The trial calculation and inversion module 200 is used to input the model constraints and initial mechanical parameters into the integrated high dam-dam foundation model for trial calculation to verify the rationality of the model. At the same time, based on the dam body displacement and stress monitoring data, a machine learning algorithm is used to carry out parameter inversion to obtain optimized mechanical parameters. The verification calculation module 300 is used to input optimized mechanical parameters into the model and calculate the dam displacement and stress if the high dam-dam foundation integrated model verification is qualified; if the model verification is unqualified, the model construction and trial calculation verification process is repeated. The error evaluation module 400 is used to obtain and analyze the dam displacement and stress data calculated by the comparative analysis model and the field monitoring data. If the error meets the preset threshold, the dam displacement and stress are predicted based on the optimized mechanical parameters. Otherwise, the parameter inversion process is re-implemented to realize the evaluation of the long-term safety of the high dam project.

[0069] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0070] Example 3 To implement the methods of the above embodiments, the present invention also provides an electronic device, which includes a memory and a processor; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the various steps of the methods described above.

[0071] Example 4 To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in the foregoing embodiments.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0073] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0074] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A long-term safety evaluation method for high dam projects based on a hybrid data and mechanism approach, characterized in that, include: S1. Based on geological survey data and high dam construction data, an integrated model of high dam and dam foundation is constructed. The constraints and initial mechanical parameters of the integrated model of high dam and dam foundation are obtained through finite element analysis. On-site monitoring of the dam body is carried out, and dam body displacement monitoring data and dam body stress monitoring data are collected. S2, Input the model constraints and initial mechanical parameters into the integrated high dam-dam foundation model for trial calculation to verify the rationality of the model. At the same time, based on the dam body displacement and stress monitoring data, use machine learning algorithms to perform parameter inversion and obtain optimized mechanical parameters. S3. If the integrated high dam-dam foundation model verification is qualified, the optimized mechanical parameters will be input into the model to calculate the dam displacement and stress. If the model verification is unqualified, the model construction and trial calculation verification process will be repeated. S4 compares and analyzes the displacement and stress data of the dam body calculated by the model with the data monitored on site, obtains and analyzes the error. If the error meets the preset threshold, the displacement and stress of the dam body are predicted based on the optimized mechanical parameters. Otherwise, the parameter inversion process is repeated to realize the evaluation of the long-term safety of the high dam project.

2. The method according to claim 1, characterized in that, The construction of the integrated high dam-dam foundation model also includes: Based on geological survey data, contour maps of the dam foundation area were obtained to determine the modeling scope and to analyze the stratigraphic structure, joint and fracture development characteristics. Modeling software was used to create the original mountain model, joint and fissure model, and dam structure model, respectively. Each sub-model is divided into grids and integrated into a unified model of high dam and dam foundation.

3. The method according to claim 2, characterized in that, The process of obtaining the constraint conditions and initial mechanical parameters of the integrated high dam-dam foundation model through finite element analysis also includes: Determine the initial conditions, boundary conditions, and external loads of the model, wherein the external loads include water pressure and sediment pressure; Physical and soil mechanics tests were conducted to obtain the initial mechanical parameters required for each model. These initial mechanical parameters included density, elastic modulus, Poisson's ratio, tensile strength, cohesion, and internal friction angle.

4. The method according to claim 3, characterized in that, Verifying the model's validity includes: The obtained initial conditions, boundary conditions, external loads, and initial mechanical parameters of the model are input into the integrated high dam-dam foundation model for trial calculation; The verification is determined by judging whether the model converges during the trial calculation. If the model does not converge, the verification is deemed unqualified; if the model converges, the verification is deemed qualified.

5. The method according to claim 4, characterized in that, The method of using machine learning algorithms to perform parameter inversion to obtain optimized mechanical parameters includes: The dam displacement and stress monitoring data are preprocessed, and the preprocessed monitoring data are used to perform parameter inversion using a preset machine learning inversion algorithm to obtain the optimized mechanical parameters required for each model. The range of values ​​for the optimized mechanical parameters is defined by the initial mechanical parameters obtained from the experiment, ensuring the rationality of the optimized mechanical parameters.

6. The method according to claim 5, characterized in that, The process for determining whether the error meets a preset threshold includes: The error judgment threshold is set at 10% of the monitoring data. The error results of the dam displacement and stress data calculated by the model and the field monitoring data are compared with this threshold. If the error result is greater than the threshold, return to the process corresponding to parameter inversion, adjust the inversion model parameters and optimize the model before re-performing parameter inversion; If the error result is less than the threshold, the subsequent displacement and stress prediction process will proceed.

7. A long-term safety evaluation device for high dam projects based on a hybrid data and mechanism approach, characterized in that, include: The modeling and monitoring module is used to construct an integrated model of the high dam and its foundation based on geological survey data and high dam construction data. It obtains the constraints and initial mechanical parameters of the integrated model of the high dam and its foundation through finite element analysis, and conducts on-site monitoring of the dam body, collecting dam body displacement monitoring data and dam body stress monitoring data. The trial calculation and inversion module is used to input the model constraints and initial mechanical parameters into the integrated high dam-dam foundation model for trial calculation to verify the rationality of the model. At the same time, based on the dam body displacement and stress monitoring data, machine learning algorithms are used to carry out parameter inversion to obtain optimized mechanical parameters. The verification calculation module is used to input optimized mechanical parameters into the model and calculate the dam displacement and stress if the high dam-dam foundation integrated model verification is qualified; if the model verification is unqualified, the model construction and trial calculation verification process is repeated. The error evaluation module is used to obtain and analyze the dam displacement and stress data calculated by the comparative analysis model and the field monitoring data. If the error meets the preset threshold, the dam displacement and stress are predicted based on the optimized mechanical parameters. Otherwise, the parameter inversion process is repeated to realize the evaluation of the long-term safety of the high dam project.

8. The apparatus according to claim 7, characterized in that, The trial inversion module is also used for: The dam displacement and stress monitoring data are preprocessed, and the preprocessed monitoring data are used to perform parameter inversion using a preset machine learning inversion algorithm to obtain the optimized mechanical parameters required for each model. The range of values ​​for the optimized mechanical parameters is defined by the initial mechanical parameters obtained from the experiment, ensuring the rationality of the optimized mechanical parameters.

9. An electronic device, characterized in that, Including processor and memory; The processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the method as described in any one of claims 1-6.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.

Images