An arch dam three-dimensional geological model and parameterized design data fusion system, method and application

Abstract

The application relates to the technical field of architectural design, in particular to an arch dam three-dimensional geological model and parameterized design data fusion system, method and application, which comprises the following modules: a data acquisition and interface module which acquires CAD data and three-dimensional geological data; a data preprocessing and feature engineering module which pre-processes the CAD data and three-dimensional geological data and extracts features; a three-dimensional model data fusion engine module which constructs a three-dimensional digital model with unified structure parameters; an analysis and evaluation module based on machine learning which uses a machine learning model to perform performance prediction, key area identification and CAE result interpretation on the three-dimensional digital model and generates performance evaluation information; and an arch dam design parameter optimization and recommendation module which uses an intelligent optimization algorithm to optimize arch dam design parameters and outputs a final design parameter combination, thereby improving the accuracy of arch dam parameterized design, the decision-making efficiency and the intelligent level of engineering design.

Description

Technical Field

[0001] This invention relates to the field of architectural design technology, specifically to a system, method, and application for fusing three-dimensional geological models and parametric design data of arch dams. Background Technology

[0002] Arch dams, as an important structural form in hydraulic engineering, are widely used in high dam construction due to their excellent mechanical properties and economic efficiency. Arch dams effectively transfer water pressure to the rock masses on both banks through their unique three-dimensional curved surface shape, significantly saving engineering materials. The design of arch dams is a complex process, involving not only the construction of a three-dimensional model of the dam body itself but also relying on the topography and geological conditions of the dam site area. Therefore, accurate three-dimensional geological models and reasonable parametric three-dimensional model design of the dam body are fundamental to ensuring the safe and economical operation of arch dams.

[0003] Traditional arch dam design processes typically rely on two-dimensional drawings, simplified computational models, and empirical formulas for design and analysis. This approach has limitations when dealing with complex three-dimensional geometries and heterogeneous, anisotropic geological conditions. First, traditional methods struggle to accurately and completely represent complex geological conditions in a three-dimensional model, and also find it difficult to effectively construct and adjust the complex three-dimensional model of the arch dam itself, leading to deviations between the designed three-dimensional model and the actual engineering geological conditions and structural morphology. Second, dam design often employs a trial-and-error approach, resulting in numerous iterations and low efficiency, making it difficult to quickly find the optimal solution based on a detailed three-dimensional model when comparing multiple options.

[0004] With the development of computer technology, computer-aided design (CAD), computer-aided engineering (CAE), and parametric design techniques have been widely applied in the engineering field. Parametric design allows for the rapid generation and adjustment of 3D structural models by modifying key parameters, improving design efficiency. Meanwhile, 3D geological modeling technology has matured, enabling the construction of detailed 3D geological models based on survey data to more realistically reflect underground structures. These 3D models provide a data foundation for engineering design.

[0005] However, in current arch dam design practice, efficiently and automatically integrating refined 3D geological model data with parametric 3D structural model design data remains a challenge. Existing design processes often have barriers between the 3D geological model construction and structural 3D model design stages, resulting in poor data transfer and low information integration. The complex boundary conditions and rock mass parameters provided by the 3D geological model are difficult to directly and seamlessly apply to the parametric structural 3D model design and optimization process. Designers need to spend a significant amount of time and effort on manual 3D model data conversion, model docking, and information integration, which is not only inefficient but also prone to introducing errors, affecting the accuracy and reliability of the final design results based on the 3D model. Therefore, how to automatically adjust the parametrically designed dam 3D model based on the 3D geological model to achieve adaptive design with geological-structural coupling is a problem that existing technologies need to effectively solve.

[0006] To address this, a system and method for fusing three-dimensional geological models and parametric design data of arch dams are proposed. Summary of the Invention

[0007] The purpose of this invention is to provide a system, method, and application for fusing three-dimensional geological models and parametric design data of arch dams. The system includes: a data acquisition and interface module for acquiring CAD data and three-dimensional geological data; a data preprocessing and feature engineering module for preprocessing and feature extraction of the CAD and three-dimensional geological data; a three-dimensional model data fusion engine module for constructing a three-dimensional digital model with unified structural parameters; a machine learning-based analysis and evaluation module for using machine learning models to predict the performance of the three-dimensional digital model, identify key areas, and interpret CAE results, generating performance evaluation information; and an arch dam design parameter optimization and recommendation module for using intelligent optimization algorithms to optimize arch dam design parameters and output the final design parameter combination, thereby improving the accuracy, decision-making efficiency, and intelligent level of arch dam parametric design.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A system for fusing three-dimensional geological models and parametric design data of arch dams includes:

[0010] The data acquisition and interface module is used to acquire CAD data of the arch dam and three-dimensional geological data of the arch dam area;

[0011] The data preprocessing and feature engineering module is used to preprocess the CAD data and the three-dimensional geological data to generate standardized three-dimensional model data, and to extract features to generate geometric features and engineering parameter features.

[0012] The 3D model data fusion engine module is used to perform spatial alignment, geometric intersection, and attribute mapping on the standardized 3D model data to generate a 3D digital model.

[0013] The machine learning-based analysis and evaluation module is used to use a pre-trained machine learning model, combined with the geometric features and the engineering parameter features, to perform engineering performance prediction, key area identification, and CAE calculation result interpretation on the three-dimensional digital model, and generate performance evaluation information.

[0014] The arch dam design parameter optimization and recommendation module is used to iteratively optimize the arch dam design parameters based on the performance evaluation information and in combination with preset optimization objectives and constraints, and output the final design parameter combination.

[0015] Furthermore, the data preprocessing process includes:

[0016] The CAD data and the three-dimensional geological data are transformed into the same engineering coordinate system by applying a preset coordinate transformation matrix to obtain preliminary three-dimensional model data;

[0017] The mesh geometry data in the preliminary 3D model data is subjected to geometry cleaning and repair operations, including mesh denoising, hole filling, non-manifold topology repair, normal unification, and removal of duplicate geometric units, to obtain the normalized 3D model data.

[0018] Furthermore, the feature extraction process includes: applying a normalization algorithm to the arch dam design parameters in the normalized three-dimensional model data to generate the engineering parameter features; calculating the geometric features from the normalized three-dimensional model data, the geometric features including: the arch dam geometric features calculated for the arch dam, including: the total volume of the arch dam, the water flow surface area at the design water level, and the average principal curvature value calculated at the preset cap beam and riverbed shoulder; and the arch dam area geometric features calculated for the arch dam area, including: the average geological thickness, and the strike parameters and dip angle parameters of the geological structural surfaces.

[0019] Furthermore, the 3D model data fusion engine module specifically includes:

[0020] The iterative nearest point algorithm is used to perform spatial alignment processing on the three-dimensional geometric model and three-dimensional geological model of the arch dam in the normalized three-dimensional model data, so as to obtain the aligned three-dimensional geometric model and the aligned three-dimensional geological model of the arch dam.

[0021] Using the aligned three-dimensional geometric model of the arch dam and the terrain data portion of the standardized three-dimensional model data, a difference operation of the three-dimensional geometric model is performed to generate a geometric model of the dam foundation excavation surface.

[0022] Using the aligned three-dimensional geometric model of the arch dam, and the geological structure part of the aligned three-dimensional geological model, the intersection operation of the three-dimensional geometric models is performed to determine the three-dimensional contact interface geometric model.

[0023] The three-dimensional contact interface geometric model, and the adjacent foundation unit jointly defined by the dam foundation excavation surface geometric model and the aligned three-dimensional geological model, are defined as the target geometric entity;

[0024] For the geological structure portion of the aligned 3D geological model, a structure query mechanism is established to query and obtain rock mass mechanical parameters and geological unit identifiers. The rock mass mechanical parameters and geological unit identifiers are assigned to the target geometric entity by executing a nearest neighbor assignment algorithm and a weighted average interpolation algorithm. The aligned 3D geometric model of the arch dam, the geometric model of the dam foundation excavation surface, the geometric model of the 3D contact interface, and the target geometric entity together constitute the 3D digital model.

[0025] Furthermore, the machine learning-based analysis and evaluation module specifically includes:

[0026] Using a pre-trained first machine learning model, combined with the geometric features and engineering parameter features, the engineering performance of the three-dimensional digital model is predicted. The engineering performance prediction includes the stress distribution, displacement deformation characteristic values, and overall stability index values ​​of the arch dam. Using a pre-trained second machine learning model, combined with the geometric features and engineering parameter features, the three-dimensional digital model is analyzed to identify key areas, including high stress concentration areas, potential cracking areas, and abnormal deformation areas. For the CAE calculation results obtained from the system interface, a pre-trained third machine learning model is applied for analysis and interpretation to extract the performance information of the design scheme.

[0027] Furthermore, the arch dam design parameter optimization and recommendation module specifically includes: using a genetic algorithm to perform iterative optimization on the arch dam design parameters and outputting the final design parameter combination; the iterative optimization is based on a preset optimization objective, which includes a first index and a second index, wherein the first index is to minimize the total volume of the arch dam concrete; and the second index is to minimize the maximum principal tensile stress of the dam body.

[0028] A method for fusing three-dimensional geological models and parametric design data of arch dams includes:

[0029] Obtain CAD data of the arch dam, as well as 3D geological data of the arch dam area;

[0030] The CAD data and 3D geological data are preprocessed to generate standardized 3D model data, and feature extraction is performed to generate geometric features and engineering parameter features.

[0031] The standardized 3D model data is spatially aligned, geometrically intersected, and attribute-mapped to generate a 3D digital model.

[0032] Using a pre-trained machine learning model, combined with the geometric features and engineering parameter features, the three-dimensional digital model is used to predict engineering performance, identify key areas, and interpret CAE calculation results to generate performance evaluation information.

[0033] Based on the performance evaluation information and combined with the preset optimization objectives and constraints, an intelligent optimization algorithm is used to iteratively optimize the arch dam design parameters and output the final combination of design parameters.

[0034] A computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the above-described method.

[0035] A computer device includes a memory, a processor, and a program stored in the memory and executable thereon, the program being executed by the processor to perform the steps of the method described above.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0037] 1. This invention achieves efficient acquisition, preprocessing, and deep fusion of arch dam CAD data and complex 3D geological data through the collaborative work of a data acquisition and interface module, a data preprocessing and feature engineering module, and a 3D model data fusion engine module. Utilizing spatial alignment, geometric intersection, and attribute mapping parameters, a standardized and unified 3D digital model that accurately reflects engineering geological conditions is constructed. This model overcomes the problem of insufficient analytical accuracy caused by scattered data sources and inconsistent models in traditional design, providing a precise and comprehensive multi-source information integration model foundation for subsequent intelligent analysis and parameter optimization.

[0038] 2. This invention achieves rapid prediction of the engineering performance of arch dam 3D digital models by using multiple pre-trained machine learning models and deeply integrating geometric features and engineering parameter features extracted from standardized 3D models. Secondly, it enables intelligent identification of key risk areas. Furthermore, it automates the interpretation of external CAE calculation results. This not only improves the efficiency and accuracy of analyzing the response of arch dam structures under complex engineering conditions but also provides forward-looking and multi-dimensional quantitative evidence for the comprehensive evaluation and scientific decision-making of design schemes.

[0039] 3. This invention combines performance evaluation information with optimization objectives, such as minimizing the total concrete volume and controlling stress and constraints in key areas. It employs intelligent algorithms, such as genetic algorithms, to automatically iteratively optimize arch dam design parameters, outputting the final optimized parameter combination. This not only effectively solves the problems of low optimization efficiency and difficulty in achieving global optimum in traditional parametric design, improving the intelligence level and decision-making efficiency of arch dam design, but also helps achieve the best balance between engineering economy and safety, promoting the development of arch dam engineering towards more refined and intelligent design. Attached Figure Description

[0040] Figure 1 A schematic diagram of the structure of the arch dam three-dimensional geological model and parametric design data fusion system provided by the present invention.

[0041] Figure 2 This invention provides a flowchart illustrating the implementation of the 3D model data fusion engine module.

[0042] Figure 3 This is a flowchart illustrating the method for fusing three-dimensional geological models and parametric design data of arch dams provided by the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0044] Please see Figures 1 to 3 This invention provides a system and method for fusing three-dimensional geological models and parametric design data of arch dams. The technical solution is as follows:

[0045] Example 1:

[0046] To meet the demands of large-scale water conservancy and hydropower projects for dam structures that are both economically efficient and highly safe, high arch dams are widely adopted due to their excellent mechanical properties and material utilization efficiency. However, the design and analysis process of high arch dams is extremely complex, requiring precise integration of complex three-dimensional geological conditions (such as irregular terrain, faults, rock mass zoning, and anisotropic material properties) and refined parametric design of the dam body. Traditional arch dam design and analysis methods, such as those based on simplified two-dimensional profiles or empirical formulas, manual transfer and conversion of geological and structural model data, and manual trial-and-error parameter optimization processes, have many limitations: they cannot achieve accurate and automated fusion of geological and structural models; they fail to fully consider the coupling effects of multi-source data and complex working conditions; they lack intelligent analysis and multi-objective collaborative optimization capabilities; and their design iteration cycles are long and inefficient.

[0047] To overcome the limitations of existing arch dam design and analysis methods, this embodiment, taking a high arch dam project of a large hydropower station as an example, elaborates on a system for fusing three-dimensional geological models and parametric design data for arch dams. Figure 1 As shown, it includes:

[0048] refer to Figure 1 The data acquisition and interface module is used to acquire CAD data of the arch dam and three-dimensional geological data of the arch dam area.

[0049] The arch dam area refers to the three-dimensional spatial range that must be covered in the evaluation of a specific arch dam project, including the dam site, the dam shoulders on both banks, the foundation affected by the load, as well as key topography and major geological structures.

[0050] Specifically, the process of acquiring CAD data for arch dams includes:

[0051] Export the STEP AP242 file from the design unit's CAD platform (AutoCAD or CATIA) and import it into the system to obtain the initial 3D geometric model. This model showcases the complex curved surface structure of the dam, covering key components such as the upstream face, downstream face, dam crest, and dam foundation, as well as the initial arch dam design parameters, such as the dam crest centerline elevation of 186.0 meters, the lowest point elevation of the dam foundation of 1.0 meter, the radius and center coordinates of the arch ring at typical elevations on the upstream face, and the control point parameters or curve relationships for the dam thickness variation.

[0052] In addition, the process of acquiring three-dimensional geological data for the arch dam area includes:

[0053] Acquire topographic data, including: collecting a 1:1000 scale digital elevation model (DEM) of the dam site area in GeoTIFF format, covering an area of ​​1 km upstream of the dam axis, 0.5 km downstream, and 0.5 km on each bank, to establish the geomorphic background and provide elevation benchmarks.

[0054] Acquiring geological structural data includes: introducing a three-dimensional geological model based on borehole data, geological sketches, and geophysical findings. For example, the borehole data covers 80 exploration boreholes at depths ranging from 50 to 250 meters, recording information such as lithology, rock quality indicators, and joint sets. The model is presented in a surface mesh format (such as GOCAD TSurf or general OBJ) and includes the three-dimensional features of major strata and faults.

[0055] Obtaining rock mass mechanics parameters includes extracting engineering geological parameters of each rock layer and structural plane from the attribute table associated with the three-dimensional geological model, such as elastic modulus, Poisson's ratio, cohesion, and internal friction angle, for mechanical analysis and performance prediction.

[0056] In this way, the system can integrate structural design data and geological environment data, providing data support for the standardized processing and fusion modeling of multi-source heterogeneous information.

[0057] refer to Figure 1 The data preprocessing and feature engineering module is used to preprocess the CAD data and 3D geological data to generate standardized 3D model data, and to extract features to generate geometric features and engineering parameter features.

[0058] Furthermore, the data preprocessing process includes:

[0059] First, the CAD data and the three-dimensional geological data are transformed to the same engineering coordinate system using a preset coordinate transformation matrix to obtain preliminary three-dimensional model data.

[0060] Specifically, the CAD data is typically constructed based on a local coordinate system established during the design phase, while the geological data may use a global geodetic coordinate system (such as WGS84 / UTM) or a local mapping coordinate system. To ensure the consistency and accuracy of the spatial location of the geometric data, the system employs a coordinate transformation method based on common control points to construct a seven-parameter transformation matrix (containing three translation components, three rotation components, and one scale factor) or a simplified three-parameter transformation matrix (containing only translation components), projecting all data uniformly onto the same engineering-specific coordinate system. This coordinate system can be established based on permanent measurement benchmarks in the dam site area and serves as the spatial reference for the entire model system. After coordinate transformation, the arch dam CAD model and the 3D geological model, originally located in different coordinate systems, are initially aligned in space, forming preliminary 3D model data.

[0061] Secondly, geometric cleaning and repair operations are performed on the mesh geometric data in the preliminary 3D model data to eliminate common geometric defects, improve data quality, and enhance the robustness of subsequent geometric calculations. The preliminary 3D model data, especially the geometric data derived from geological modeling or CAD model conversion, often exists in the form of mesh geometric data. This mesh geometric data is a discretized digital model that defines 3D shapes using vertices, edges, and faces, used to represent geological interfaces and structural entities. In this embodiment, the geometric cleaning and repair operations specifically include:

[0062] Mesh denoising includes: using the Laplacian smoothing algorithm or similar filtering techniques to eliminate high-frequency noise and minute disturbances on the mesh surface, making the model surface smoother and more continuous;

[0063] Hole filling includes: based on boundary loop detection and surface fitting algorithms, identifying and automatically repairing open areas in the model to ensure mesh closure;

[0064] Non-manifold topology repair includes: detecting and correcting non-manifold structures, such as when an edge is shared by multiple faces or when there are multiple disconnected regions at a vertex, and using local reconstruction or partitioning techniques to handle the situation.

[0065] Normal unification includes: calculating the direction of the normal vector of the surface patch and unifying it to point to the outside of the model or a specified direction to meet the consistency requirements of subsequent Boolean operations or contact analysis;

[0066] Repetitive geometric element removal includes detecting and removing duplicate vertices (vertices with the same spatial location or within a very small tolerance range), collinear edges, or overlapping and coplanar small elements in the model to simplify data and reduce potential computational errors.

[0067] After geometric cleaning and repair operations, standardized 3D model data with valid topology, complete geometric features, and unified data structure is obtained. In this embodiment, it includes a clearly defined set of digital data containing standardized 3D geometric models of arch dams and standardized 3D geological models (integrating topographic surface and internal geological structure information).

[0068] By performing coordinate system transformation and geometric cleaning and repair operations on the original CAD data and 3D geological data, the consistency of the 3D model data input into subsequent modules in terms of spatial reference, the integrity and accuracy of geometric morphology, and the validity of topological structure are ensured. Furthermore, the level and quality of data standardization are significantly improved, providing a reliable data foundation for subsequent 3D model fusion, feature extraction, and machine learning analysis. This avoids calculation errors or distortion of analysis results caused by problems with the quality of the original data, thereby improving the accuracy of parametric design of arch dams, decision-making efficiency, and the level of intelligence in engineering design.

[0069] Furthermore, the feature extraction process includes:

[0070] First, a normalization algorithm is applied to the arch dam design parameters (such as the actual dam height of 185m, the thickness of each control point, etc.) in the normalized three-dimensional model data to map them into the [0,1] interval, generating the engineering parameter features with consistent dimensions and stable values ​​for use by subsequent machine learning models.

[0071] Secondly, the geometric features are calculated from the normalized 3D model data, the geometric features including:

[0072] The geometric characteristics of the arch dam calculated for the arch dam include: the total volume of the arch dam (e.g., a preliminary estimate of 2.8 million cubic meters); the water-passing surface area at the design water level, such as the surface area of ​​the upstream face of the arch dam in contact with the water body under the design water level of 175m, reflecting the stress and fluid boundary conditions; and the average principal curvature value calculated at the preset capping beam and riverbed shoulder, such as at preset key locations, such as the center point of the capping beam at the top of the dam, the center point of the capping beam at 1 / 2 dam height, and the center point of the dam toe, calculating the average principal curvature value of the surface at their respective locations, used to quantify the local geometric shape and structural flexibility of the curved surface.

[0073] The geometric features of the arch dam region calculated for the arch dam region include: the average geological thickness (e.g., microcrystalline granite) that interacts with the arch dam foundation, and the strike and dip parameters of geological structural surfaces, such as the average strike (e.g., 30°) and average dip (e.g., 75°) of the main faults F1 and F2, used to describe the regional tectonic environment and the distribution of potential weak surfaces. These features together constitute the geometric feature vector.

[0074] By normalizing the original arch dam design parameters to generate engineering parameter features, and accurately calculating them from standardized 3D model data, geometric features characterizing the arch dam morphology and regional geological conditions are obtained. This provides dimensionally consistent and numerically stable input features for subsequent machine learning models, enhancing their training efficiency, prediction accuracy, and generalization ability. This makes engineering performance prediction and key area identification based on these features more reliable and accurate, thereby improving the accuracy of arch dam parametric design, decision-making efficiency, and the level of intelligence in engineering design.

[0075] refer to Figure 1 The three-dimensional model data fusion engine module is used to perform spatial alignment, geometric intersection and attribute mapping on the standardized three-dimensional model data to generate a three-dimensional digital model.

[0076] Furthermore, Figure 2 This document provides a flowchart illustrating the implementation of the 3D model data fusion engine module for this invention. (For example...) Figure 2 As shown, the 3D model data fusion engine module specifically includes:

[0077] First, the Iterative Closest Point (ICP) algorithm is used to perform spatial alignment processing on the three-dimensional geometric model and three-dimensional geological model of the arch dam in the normalized three-dimensional model data, so as to obtain the aligned three-dimensional geometric model and the aligned three-dimensional geological model of the arch dam.

[0078] Using the aligned three-dimensional geometric model of the arch dam and the terrain data portion of the standardized three-dimensional model data, a difference operation of the three-dimensional geometric model is performed to generate a geometric model of the dam foundation excavation surface; wherein, the terrain data portion refers to a three-dimensional digital surface model that accurately describes the natural undulation of the surface of the arch dam area after coordinate unification and geometric cleanup.

[0079] Using the aligned 3D geometric model of the arch dam, the intersection operation of the 3D geometric model and the geological structure part of the aligned 3D geological model is performed to determine the 3D contact interface geometric model. The geological structure part refers to the 3D model information that has also been preprocessed, characterizing the spatial distribution and geometric shape of various underground and exposed geological bodies (such as rock strata and faults) in the arch dam area, and has been associated with their corresponding rock mass mechanical parameters and geological unit identifiers.

[0080] Based on this, the three-dimensional contact interface geometric model, as well as the adjacent foundation units (e.g., hexahedral or tetrahedral units extending 50 to 100 m in depth within the influence range) jointly defined by the dam foundation excavation surface geometric model and the aligned three-dimensional geological model, are defined as target geometric entities. These target geometric entities represent the dam foundation and its surrounding key rock mass areas, and are the core spatial objects for dam stability analysis.

[0081] To ensure that the constructed 3D digital model accurately reflects engineering geological conditions, geological attributes (rock mechanics parameters and geological unit identification information) need to be assigned to the target geometric entity. To this end, the system establishes a structure query mechanism based on the Octree spatial index structure, supporting the rapid location of the geological body at any spatial point coordinate and the extraction of corresponding attribute data.

[0082] Specifically, for each target geometric entity (whether it is a surface element or a volume element), its centroid coordinates or multiple integration points are used as a reference to invoke the structural query mechanism to retrieve and obtain the corresponding geological attributes from the aligned 3D geological model, such as:

[0083] Geological unit identification information: such as microcrystalline granite, fault fracture zone;

[0084] Rock mass mechanics parameters: such as elastic modulus (e.g., 60 GPa), Poisson's ratio (e.g., 0.20), cohesion (e.g., 2.5 MPa), internal friction angle (e.g., 48°), etc.

[0085] The obtained rock mass mechanical parameters and geological unit identification information are used to assign corresponding target geometric entities through a combination or selection of the following two algorithms:

[0086] The rock mass mechanical parameters and the geological unit identifier are assigned to the target geometric entity using a nearest neighbor assignment algorithm. This includes cases where the attributes are relatively uniform or the unit falls entirely within a geological body, where the geological body attributes of the target geometric entity (or its centroid) are directly assigned to that unit.

[0087] Furthermore, the target geometric entity is assigned an attribute by performing a weighted average interpolation algorithm. This includes assigning attributes to units located near the boundaries of different geological bodies where attributes may gradually change, or to achieve a smoother transition of attributes, by using inverse distance weighting and other methods to perform attribute interpolation calculations based on the distance between the target unit and multiple data points of different attributes around it (such as measured parameter points in boreholes or boundaries of different geological bodies).

[0088] The aligned three-dimensional geometric model of the arch dam, the geometric model of the dam foundation excavation surface, the geometric model of the three-dimensional contact interface, and the target geometric entity together constitute the three-dimensional digital model.

[0089] By employing the ICP algorithm for precise spatial alignment, combined with the difference and intersection operations of the three-dimensional geometric model, and utilizing the query mechanism based on the spatial index data structure, as well as the nearest neighbor assignment and weighted average interpolation algorithm, the accurate mapping of geological attributes to the target geometric entity is achieved. This provides a high-quality, high-fidelity model foundation for subsequent refined analysis and machine learning evaluation, thereby improving the accuracy of arch dam parametric design, decision-making efficiency, and the level of intelligence in engineering design.

[0090] refer to Figure 1 The machine learning-based analysis and evaluation module is used to utilize a pre-trained machine learning model, combined with the geometric features and engineering parameter features, to perform engineering performance prediction, key area identification, and CAE calculation result interpretation on the three-dimensional digital model, and generate performance evaluation information.

[0091] Furthermore, the machine learning-based analysis and evaluation module specifically includes:

[0092] Using a pre-trained first machine learning model, and combining the geometric features and the engineering parameter features, the engineering performance of the three-dimensional digital model is predicted. The engineering performance prediction includes the stress distribution, displacement deformation feature values, and overall stability index values ​​of the arch dam.

[0093] In this embodiment, a pre-trained deep neural network (DNN) is used as the first machine learning model. The engineering parameters and some key geometric features of the current 185m high arch dam design are input to quickly predict the key performance indicators under the combined action of hydrostatic pressure at the design water level of 175m and temperature load considering temperature control measures: maximum principal tensile stress (predicted value, such as 2.8MPa), maximum principal compressive stress, maximum displacement of the dam crest along the river (such as 35mm), and overall anti-sliding stability safety factor, etc.

[0094] Using a pre-trained second machine learning model, and combining the geometric features and the engineering parameter features, the three-dimensional digital model is analyzed to identify the key regions, which include high stress concentration regions, potential cracking regions, and abnormal deformation regions.

[0095] In this embodiment, a pre-trained convolutional neural network (CNN) is used as a second machine learning model to analyze the stress calculation results or geometric features of the three-dimensional digital model, and automatically identify potential high stress concentration areas (such as dam shoulders, dam foundation heels, and areas near gallery entrances), potential cracking areas, and abnormal deformation areas.

[0096] The CAE calculation results obtained from the system interface are analyzed and interpreted using a pre-trained third machine learning model to extract performance information of the design scheme.

[0097] In this embodiment, a detailed CAE calculation result file for the current design scheme is imported. A pre-trained third machine learning model (e.g., a pattern recognition-based classification / regression model) is applied to intelligently post-process and interpret the output data (such as stress and strain values), automatically extracting information such as statistics of over-limit stress areas and preliminary judgment of crack development trends to form structured design scheme performance information.

[0098] These predicted values, identified regional information, and design scheme performance information together constitute the performance evaluation information for the current design scheme.

[0099] By utilizing different types of pre-trained machine learning models, rapid and intelligent multi-dimensional analysis and evaluation of the fused 3D digital model and related data can be performed, improving evaluation efficiency and thus enhancing the accuracy of arch dam parametric design, decision-making efficiency, and the level of intelligence in engineering design.

[0100] refer to Figure 1 The arch dam design parameter optimization and recommendation module is used to iteratively optimize the arch dam design parameters based on the performance evaluation information and in combination with preset optimization objectives and constraints, and output the final design parameter combination.

[0101] The constraints include: the maximum tensile stress in each part of the dam body shall not exceed the upper limit of the material design strength; the maximum compressive stress shall not exceed the allowable value; the horizontal displacement of the dam crest shall not exceed the specified limit; and the overall anti-sliding stability safety factor K shall not be less than the set lower limit.

[0102] Furthermore, under the premise of satisfying the above constraints, the arch dam design parameter optimization and recommendation module specifically includes: using a genetic algorithm to perform iterative optimization of the arch dam design parameters, and outputting the final design parameter combinations, which represent the better design schemes under the current optimization objectives and constraints. For example, outputting a parameter combination that can still satisfy all safety constraints while reducing the volume by 5%.

[0103] The iterative optimization is based on a preset optimization objective, which includes a first indicator and a second indicator. The first indicator is to minimize the total volume of concrete in the arch dam to reduce construction costs and material consumption. The second indicator is to minimize the maximum principal tensile stress in the dam body to improve the overall safety margin of the structure.

[0104] In this embodiment, a genetic algorithm is used as the core optimization strategy. It performs a global search of the design space using real-number encoded design parameters (such as dam height, thickness distribution, and arch curvature) combined with selection, crossover, and mutation operations. In each iteration, the algorithm scores the fitness of candidate parameter combinations and generates the next generation of candidate solutions accordingly. Through continuous iterative updates, the system ultimately outputs a set of design parameter combinations that satisfy all constraints and perform well on the optimization objective. For example, the system can recommend an optimized design scheme that reduces concrete usage by approximately 5% and controls the principal tensile stress within 2.3 MPa, thereby achieving a dual improvement in economy and safety.

[0105] By combining engineering goal-driven optimization logic with the search capabilities of intelligent algorithms, it can efficiently identify optimal or near-optimal solutions in complex design spaces, thereby improving the accuracy, decision-making efficiency, and intelligence level of arch dam parametric design.

[0106] Example 2:

[0107] To further verify the effectiveness of the present invention, based on Embodiment 1, a method for fusing three-dimensional geological models and parametric design data of arch dams is proposed, such as... Figure 3 As shown, it includes:

[0108] Obtain CAD data of the arch dam, as well as 3D geological data of the arch dam area;

[0109] The CAD data and 3D geological data are preprocessed to generate standardized 3D model data, and feature extraction is performed to generate geometric features and engineering parameter features.

[0110] The standardized 3D model data is spatially aligned, geometrically intersected, and attribute-mapped to generate a 3D digital model.

[0111] Using a pre-trained machine learning model, combined with the geometric features and engineering parameter features, the three-dimensional digital model is used to predict engineering performance, identify key areas, and interpret CAE calculation results, generating performance evaluation information.

[0112] To verify the performance of this invention in automatically identifying key regions of structural response, a comparative experiment was conducted on the automatic identification of high-stress regions in the stress analysis results of arch dams. In the experiment, the key region identification model based on a three-dimensional digital model using a convolutional neural network (CNN) was compared with a common baseline image processing algorithm (a CNN model based on raw CAD data and three-dimensional geological data). Evaluation metrics included accuracy, precision, recall, F1 score, and average processing time per image for high-stress region identification. The comparison results are shown in Table 1. The results demonstrate that this invention outperforms the traditional CNN algorithm in all performance metrics for automatically identifying high-stress regions in arch dams.

[0113] Table 1 Performance Comparison of Key Region Identification Algorithms

[0114]

[0115] Based on the performance evaluation information and combined with the preset optimization objectives and constraints, an intelligent optimization algorithm is used to iteratively optimize the arch dam design parameters and output the final combination of design parameters.

[0116] To further verify the practical application effect of this invention, a comparative experiment was conducted on the performance of optimized design parameter combinations for a high arch dam project of a large hydropower station (dam height 185m, design water level 175m). Specific constraints included: the maximum principal tensile stress in key parts of the arch dam must not exceed 2.5MPa, the maximum principal compressive stress must not exceed 20MPa, and the overall anti-sliding stability safety factor must be greater than or equal to 1.3.

[0117] The method of this invention was compared with a traditional baseline method. This traditional baseline method employs a standard genetic algorithm, but in each iteration, it directly calls the complete Computer-Aided Engineering (CAE) simulation process (including 3D modeling, mesh generation, finite element calculation, and post-processing) to obtain the performance indicators of the design scheme when evaluating the fitness of individuals. To ensure the fairness of the comparison, both methods started with the same randomly generated initial design parameter population and performed a total of 200 iterations.

[0118] Table 2 Comparison of Optimization Performance of Final Design Parameter Combinations

[0119]

[0120] As shown in Table 2, the present invention demonstrates better performance in terms of both the quality of optimization results and optimization efficiency in optimizing arch dam design parameters.

[0121] A computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the above-described method.

[0122] A computer device includes a memory, a processor, and a program stored in the memory and executable thereon, the program being executed by the processor to perform the steps of the method described above.

[0123] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented, in whole or in part, as a computer program product, the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape) or an optical medium.

[0124] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A system for fusing three-dimensional geological models and parametric design data of arch dams, characterized in that, include: The data acquisition and interface module is used to acquire CAD data of the arch dam and three-dimensional geological data of the arch dam area; The data preprocessing and feature engineering module is used to preprocess the CAD data and the three-dimensional geological data to generate standardized three-dimensional model data, and to extract features to generate geometric features and engineering parameter features. The 3D model data fusion engine module is used to perform spatial alignment, geometric intersection, and attribute mapping on the standardized 3D model data to generate a 3D digital model. The machine learning-based analysis and evaluation module is used to use a pre-trained machine learning model, combined with the geometric features and the engineering parameter features, to perform engineering performance prediction, key area identification, and CAE calculation result interpretation on the three-dimensional digital model, and generate performance evaluation information. The arch dam design parameter optimization and recommendation module is used to iteratively optimize the arch dam design parameters based on the performance evaluation information and in combination with preset optimization objectives and constraints, and output the final design parameter combination.

2. The arch dam three-dimensional geological model and parametric design data fusion system according to claim 1, characterized in that, The data preprocessing process includes: The CAD data and the three-dimensional geological data are transformed into the same engineering coordinate system by applying a preset coordinate transformation matrix to obtain preliminary three-dimensional model data; The mesh geometry data in the preliminary 3D model data is subjected to geometry cleaning and repair operations, including mesh denoising, hole filling, non-manifold topology repair, normal unification, and removal of duplicate geometric units, to obtain the normalized 3D model data.

3. The arch dam three-dimensional geological model and parametric design data fusion system according to claim 1, characterized in that, The feature extraction process includes: applying a normalization algorithm to the arch dam design parameters in the normalized 3D model data to generate the engineering parameter features; calculating the geometric features from the normalized 3D model data, the geometric features including: the arch dam geometric features calculated for the arch dam, including: the total volume of the arch dam, the water flow surface area, and the average principal curvature value calculated at the preset cap beam and riverbed shoulder; and the arch dam area geometric features calculated for the arch dam area, including: the average geological thickness, and the strike parameters and dip angle parameters of the geological structural surfaces.

4. The arch dam three-dimensional geological model and parametric design data fusion system according to claim 1, characterized in that, The 3D model data fusion engine module specifically includes: The iterative nearest point algorithm is used to perform spatial alignment processing on the three-dimensional geometric model and three-dimensional geological model of the arch dam in the normalized three-dimensional model data, so as to obtain the aligned three-dimensional geometric model and the aligned three-dimensional geological model of the arch dam. Using the aligned three-dimensional geometric model of the arch dam and the terrain data portion of the standardized three-dimensional model data, a difference operation of the three-dimensional geometric model is performed to generate a geometric model of the dam foundation excavation surface. Using the aligned three-dimensional geometric model of the arch dam, and the geological structure part of the aligned three-dimensional geological model, the intersection operation of the three-dimensional geometric models is performed to determine the three-dimensional contact interface geometric model. The three-dimensional contact interface geometric model, and the adjacent foundation unit jointly defined by the dam foundation excavation surface geometric model and the aligned three-dimensional geological model, are defined as the target geometric entity; For the geological structure portion of the aligned 3D geological model, a structure query mechanism is established to query and obtain rock mass mechanical parameters and geological unit identifiers. The rock mass mechanical parameters and geological unit identifiers are then assigned to the target geometric entity by executing a nearest neighbor assignment algorithm and a weighted average interpolation algorithm. The aligned 3D geometric model of the arch dam, the geometric model of the dam foundation excavation surface, the 3D contact interface geometric model, and the target geometric entity together constitute the 3D digital model.

5. The arch dam three-dimensional geological model and parametric design data fusion system according to claim 1, characterized in that, The machine learning-based analysis and evaluation module specifically includes: Using a pre-trained first machine learning model, combined with the geometric features and engineering parameter features, the engineering performance of the three-dimensional digital model is predicted. The engineering performance prediction includes the stress distribution, displacement deformation characteristic values, and overall stability index values ​​of the arch dam. Using a pre-trained second machine learning model, combined with the geometric features and engineering parameter features, the three-dimensional digital model is analyzed to identify key areas, including high stress concentration areas, potential cracking areas, and abnormal deformation areas. For the CAE calculation results obtained from the system interface, a pre-trained third machine learning model is applied for analysis and interpretation to extract the performance information of the design scheme.

6. The arch dam three-dimensional geological model and parametric design data fusion system according to claim 1, characterized in that, The arch dam design parameter optimization and recommendation module specifically includes: using a genetic algorithm to perform iterative optimization on the arch dam design parameters and outputting the final design parameter combination; the iterative optimization is based on a preset optimization objective, which includes a first indicator and a second indicator, wherein the first indicator is to minimize the total volume of the arch dam concrete; and the second indicator is to minimize the maximum principal tensile stress of the dam body.

7. A method for fusing three-dimensional geological models and parametric design data of arch dams, characterized in that, include: Obtain CAD data of the arch dam, as well as 3D geological data of the arch dam area; The CAD data and 3D geological data are preprocessed to generate standardized 3D model data, and feature extraction is performed to generate geometric features and engineering parameter features. The standardized 3D model data is spatially aligned, geometrically intersected, and attribute-mapped to generate a 3D digital model. Using a pre-trained machine learning model, combined with the geometric features and engineering parameter features, the three-dimensional digital model is used to predict engineering performance, identify key areas, and interpret CAE calculation results to generate performance evaluation information. Based on the performance evaluation information and combined with the preset optimization objectives and constraints, an intelligent optimization algorithm is used to iteratively optimize the arch dam design parameters and output the final combination of design parameters.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the method of claim 7.

9. A computer device, characterized in that, The computer device includes a memory, a processor, and a program stored in and executable on the memory, the program being executed by the processor to perform the steps of the method as described in claim 7.

Images