Three-dimensional design method, system and equipment for water conservancy project and storage medium

Abstract

The invention provides a hydraulic engineering three-dimensional design method, system and equipment and a storage medium, and relates to the field of three-dimensional design. The method comprises the steps of obtaining topographic data, hydrological data and engineering design requirements of an area where a project is located; performing feature extraction on the topographic data and the hydrological data to obtain key feature parameters; determining design constraint conditions according to engineering design requirements, and establishing a three-dimensional design optimization model based on the key feature parameters as design variables and the design constraint conditions; solving the three-dimensional design optimization model through a genetic algorithm to obtain a group of non-inferior solutions as a first candidate design scheme; optimizing the first candidate design scheme through a preset reinforcement learning model to obtain a second candidate design scheme; and performing simulation analysis based on the second candidate design scheme to obtain feedback information, and optimizing the second candidate design scheme based on the feedback information to obtain a target design scheme. Through the scheme, the hydraulic engineering design efficiency and the optimization degree of the design scheme are improved.

Description

Technical Field

[0001] The present application relates to the field of three-dimensional design, and specifically to a three-dimensional design method, system, equipment and storage medium for water conservancy projects. Background Art

[0002] Water conservancy project design is a complex system engineering, involving multiple factors such as topography, hydrology, structure, and materials. Traditional water conservancy project design mainly relies on the experience of designers, and gradually optimizes the design plan through trial and error and manual adjustment. This method has the problems of low efficiency and insufficient optimization. Summary of the invention

[0003] The present application provides a three-dimensional design method, system, equipment and storage medium for a water conservancy project, which can improve the design efficiency of the water conservancy project and the optimization degree of the design scheme.

[0004] In a first aspect, the present application provides a three-dimensional design method for a water conservancy project, the method comprising: Obtain topographic data, hydrological data and engineering design requirements for the project area; Extract features from terrain data and hydrological data to obtain key feature parameters; Determine design constraints according to engineering design requirements, and establish a three-dimensional design optimization model based on key characteristic parameters as design variables, engineering design requirements and design constraints; The three-dimensional design optimization model is solved by genetic algorithm to obtain a set of non-inferior solutions as the first candidate design scheme; The first candidate design scheme is optimized by using a preset reinforcement learning model to obtain a second candidate design scheme; A simulation analysis is performed based on the second candidate design solution to obtain feedback information, and the second candidate design solution is optimized based on the feedback information to obtain a target design solution.

[0005] By adopting the above technical scheme, this application obtains the topographic data, hydrological data and engineering design requirements of the project area, uses feature extraction technology to extract key feature parameters, and determines the design constraints in combination with the engineering design requirements, and builds a three-dimensional design optimization model on this basis. Since the model fully considers the topographic and hydrological characteristics of the project area, it can comprehensively depict the actual situation of the project, and the design variables, constraints and engineering requirements are closely related, and the optimization objectives cover various performance indicators, it can accurately and objectively describe the design problem, laying a solid foundation for subsequent optimization. Furthermore, by using the genetic algorithm, a global optimization technology, to solve the model, comprehensively weighing various objectives, a set of balanced and feasible non-inferior solutions is obtained as the first candidate design scheme, and preliminary macro optimization is achieved in the design space. Next, the preset reinforcement learning model is innovatively introduced. On the basis of the preliminary scheme, local features are gradually improved through continuous trial and error and self-reinforcement, and finally a more refined second candidate design scheme is obtained, further tapping the design potential. Finally, a simulation analysis was carried out on the second candidate design scheme. Various performance indicators were evaluated with high fidelity through numerical simulation, and the simulation results were used as feedback information to feed back into the optimization process, continuously guiding parameter adjustments. Through multiple iterations, the optimal solution was continuously approached, and finally a target design scheme that met multiple requirements and had outstanding performance was obtained.

[0006] In a second aspect of the present application, a three-dimensional design method system for a water conservancy project is provided, comprising: Data acquisition module, used to obtain topographic data, hydrological data and engineering design requirements of the project area; Feature extraction module, used to extract features from terrain data and hydrological data to obtain key feature parameters; A model building module is used to determine design constraints according to engineering design requirements and to build a three-dimensional design optimization model based on key characteristic parameters as design variables, engineering design requirements and design constraints; A solution module is used to solve the three-dimensional design optimization model by using a genetic algorithm to obtain a set of non-inferior solutions as the first candidate design solution; A first optimization module, used to optimize the first candidate design solution through a preset reinforcement learning model to obtain a second candidate design solution; The second optimization module is used to perform simulation analysis based on the second candidate design solution to obtain feedback information, and optimize the second candidate design solution based on the feedback information to obtain a target design solution.

[0007] In a third aspect of the present application, a computer storage medium is provided, wherein the computer storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the above method steps.

[0008] In the fourth aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the above method.

[0009] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. This application obtains the topographic data, hydrological data and engineering design requirements of the project area, uses feature extraction technology to refine key feature parameters, and determines the design constraints in combination with the engineering design requirements, and builds a three-dimensional design optimization model on this basis. Since the model fully considers the topographic and hydrological characteristics of the project area, it can fully describe the actual situation of the project, and the design variables and constraints are closely related to the engineering requirements. The optimization goals cover various performance indicators, so it can accurately and objectively describe the design problems, laying a solid foundation for subsequent optimization.

[0010] 2. This application uses a global optimization technology called genetic algorithm to solve the model, comprehensively weighs various objectives, and obtains a set of balanced and feasible non-inferior solutions as the first candidate design scheme, achieving preliminary macro optimization in the design space. Next, the preset reinforcement learning model is innovatively introduced. On the basis of the preliminary scheme, local features are gradually improved through continuous trial and error and self-reinforcement, and finally a more refined second candidate design scheme is obtained, further tapping the design potential.

[0011] 3. This application conducts simulation analysis on the second candidate design scheme, evaluates various performance indicators with high fidelity through numerical simulation, and uses the simulation results as feedback information to feed back the optimization process, continuously guide parameter adjustment, and iterate multiple times to continuously approach the optimal solution, and finally obtain a target design scheme that meets multiple requirements and has outstanding performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic diagram of a flow chart of a three-dimensional design method for a water conservancy project provided in an embodiment of the present application; Figure 2 An architectural diagram of a three-dimensional design system for water conservancy projects provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION

[0013] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0014] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0015] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0016] In order to facilitate understanding of the method and system provided by the embodiments of the present application, before introducing the embodiments of the present application, the background of the embodiments of the present application is first introduced.

[0017] Water conservancy project design is a complex system engineering involving multiple disciplines and fields, which requires comprehensive consideration of many factors such as the terrain conditions, hydrological characteristics, structural layout, material selection, etc. in the area where the project is located. The traditional water conservancy project design process mainly relies on the personal experience and professional knowledge of designers, and gradually improves and optimizes the design plan through a series of repeated operations such as continuous trial calculations, comparisons, and corrections. Although this design method based on manual experience can guarantee the feasibility and safety of the design plan to a certain extent, it also has obvious limitations and shortcomings.

[0018] On the one hand, the efficiency of traditional design methods is low and the degree of optimization is limited. Due to the large number of parameters and complex constraints involved in water conservancy project design, it is difficult for designers to exhaustively enumerate and compare all possible design schemes within a limited time. They are often limited to a few alternatives and it is difficult to achieve the global optimal solution. At the same time, the experience and knowledge levels of designers vary, and their understanding and judgment of design problems are also subjective and blind, which limits the degree of optimization of the design scheme and makes it difficult to achieve the theoretical optimal solution.

[0019] On the other hand, traditional design methods are difficult to fully explore and utilize the specific endowments and resource advantages of the project area. Water conservancy projects often need to be adapted to local conditions. According to the characteristics of the natural conditions such as topography, geology, hydrology, and meteorology in the project area, the project structure should be reasonably arranged and the engineering materials should be optimized to achieve the purpose of adapting to local conditions and making the best use of resources. However, traditional design methods are subject to the knowledge limitations and experience blind spots of designers, and it is often difficult to fully consider and systematically analyze regional characteristics, and it is impossible to fully explore and utilize the favorable terrain conditions and resource endowments in the project area, resulting in the lack of pertinence and adaptability of the design plan, and it is difficult to maximize the project benefits.

[0020] In addition, traditional design methods also have the risk of being out of touch with reality. Since the design process mainly relies on manual experience and lacks strict, quantitative demonstration and verification methods, the design plan may not be completely consistent with the actual engineering conditions. Once some seemingly reasonable design plans are implemented, they may expose problems such as difficult construction, difficult operation and maintenance, and many safety hazards, which will affect the progress, quality and benefits of the project and even cause major safety accidents and economic losses.

[0021] After the background introduction of the above content, those skilled in the art can understand the problems existing in the prior art. The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0022] Based on the above background technology, please refer to Figure 1 , Figure 1 A flow chart of a method for three-dimensional design of a water conservancy project provided in an embodiment of the present application. The system can be implemented by a computer program or can be run as an independent tool application. Specifically, in an embodiment of the present application, the method can be applied on a server, but can also be applied to electronic devices such as a server. A method for three-dimensional design of a water conservancy project includes the following steps: S101, obtain topographic data, hydrological data and engineering design requirements of the project area; Specifically, in a three-dimensional design optimization method for a water conservancy project in the present application, the first step is to obtain the terrain data, hydrological data and engineering design requirements of the area where the project is located. The reason for obtaining these three types of data is that they are the basis and prerequisite for carrying out the design optimization of water conservancy projects. The terrain conditions and hydrological characteristics of the area where the project is located are key factors affecting multiple design links such as project layout, structural form, and construction plan, while the engineering design requirements are an important basis for determining design goals, constraints, and evaluation indicators. Only by fully and accurately mastering these three types of data can we construct an optimization model that is compatible with actual engineering conditions and obtain a practical and economically reasonable design plan.

[0023] Specifically, terrain data can be obtained through field surveys, remote sensing image interpretation, digital elevation models and other means. Through field surveys, basic data such as topographic maps and contour maps of the project area can be obtained, and key factors affecting the project layout such as terrain characteristics and geomorphic units can be inspected on site. Using high-resolution remote sensing image data, ground feature information can be extracted through manual or automatic interpretation methods, digital terrain maps can be generated, and terrain feature analysis can be further carried out. In addition, based on digital elevation models, three-dimensional terrain scenes can be quickly generated to intuitively display micro-geomorphic features such as terrain undulations, slopes and directions in the project area, providing a reference for optimizing project layout and earthwork allocation.

[0024] Hydrological data are mainly obtained through long-term observation data of hydrological stations, hydrogeological surveys, basin hydrological model simulations, etc. By collecting and collating observation data such as precipitation, river runoff, and flood levels from hydrological stations over the years, the hydrological characteristics of the project area, such as the distribution of wet and dry seasons, flood frequency, and design flood levels, can be analyzed to provide a basis for determining the scale and standards of the project. Through hydrogeological surveys, the groundwater storage conditions, recharge and discharge characteristics, and water level dynamics in the project area and its upstream and downstream areas can be mastered to provide support for demonstrating the mutual influence of groundwater and the project. With the help of the basin hydrological model, the runoff regulation effect after the implementation of the project can be simulated, and the water balance and supply and demand contradictions during the operation of the project can be analyzed.

[0025] The main methods of obtaining engineering design requirements are on-site surveys, fund collection and review, and special research. Through on-site field surveys, we can have an in-depth understanding of the engineering tasks and goals, functional positioning, construction conditions, etc., and fully communicate with local management departments and beneficiaries, and listen to opinions and suggestions from all parties. By widely collecting and reviewing various types of literature and materials, such as national and industry-related specifications, special plans, and existing engineering operation data, we can accurately grasp the binding requirements such as design standards, construction procedures, and technical specifications. In view of the different functions of the project, such as flood control, water supply, and power generation, special research is also needed to demonstrate the rationality of the project scale, standards, and tasks, and to refine and quantify various design indicators.

[0026] S102, extracting features from terrain data and hydrological data to obtain key feature parameters; Specifically, after obtaining the topographic data and hydrological data of the project area, the next step is to extract features from these data to obtain key feature parameters. The reason for feature extraction is that the original topographic data and hydrological data often have the characteristics of large data volume, information redundancy, and diverse formats, which are difficult to directly apply to the construction and solution of design optimization models. Through feature extraction, key information that has a significant impact and restrictive effect on engineering design optimization can be mined from the original data, and it can be quantified, structured, and parameterized, thereby greatly reducing the complexity and computational complexity of the optimization problem and improving the efficiency and accuracy of the optimization.

[0027] Specifically, the feature extraction of terrain data can be done by digital terrain analysis, quantitative description of landforms and other methods. Using spatial analysis tools such as GIS, various terrain feature parameters such as elevation, slope, slope aspect, undulation, terrain roughness, etc. can be extracted from the digital elevation model, and the spatial distribution characteristics of the terrain can be characterized by terrain profile analysis, terrain statistical analysis and other means. For typical geomorphic units such as river valleys, mesas, terraces, etc. that are closely related to the project layout, their geometric feature parameters such as boundaries, morphology, connectivity, etc. can also be extracted. In addition, for different landform types such as mountainous areas, hills, and plains where water conservancy projects are located, their respective feature parameter systems can be summarized and summarized, and a terrain feature parameter library can be established to provide a reference for engineering design.

[0028] The feature extraction of hydrological data mainly adopts methods such as hydrological frequency analysis and hydrological time series analysis. By statistically analyzing the hydrological observation data for many years, a series of design hydrological parameters such as design flood, design low water, and multi-year average runoff can be obtained to describe the variation law of hydrological elements in the project area. Using the hydrological frequency calculation and parameter estimation theory, the design flood and design rainfall of different recurrence periods can be deduced to provide a basis for the demonstration of project scale and standards. The analysis of hydrological time series can reveal the periodicity, trend, and mutation characteristics of runoff and water level, distinguish the dry and wet years, the wet and dry cycles, predict the long-term hydrological change trend, and guide the project planning. In the process of extracting hydrological parameters, we pay attention to the role of hydrological models, use the basin rainfall runoff model, reservoir scheduling model, etc., analyze the runoff process and reservoir capacity changes after the implementation of the project through scenario simulation, and extract the key parameters reflecting the hydrological response characteristics.

[0029] Based on the above embodiment, as an optional embodiment, feature extraction is performed on the terrain data and the hydrological data to obtain key feature parameters, including: S201, converting terrain data into a terrain digital elevation model, and performing feature extraction based on the terrain digital elevation model to obtain terrain feature parameters; Specifically, when conducting simulation analysis based on the second candidate design scheme, we must first select appropriate simulation models and parameters according to different simulation purposes and requirements. For the structural safety analysis of hydraulic structures, the finite element method can be used to establish a three-dimensional stress-strain model of dams, spillways, and water diversion structures, simulate their stress-deformation characteristics under loads such as water flow, earthquakes, and temperature, and examine safety indicators such as stress distribution, displacement trend, and crack development. For the hydrological and hydraulic analysis of the reservoir operation process, the numerical simulation method can be used to establish a three-dimensional hydrodynamic model of the reservoir area, simulate its hydraulic characteristics such as water level changes, flow velocity distribution, and silt deposition under different inflows and different operation and scheduling modes, and analyze the comprehensive benefits of cascade development, flood control regulation, water supply and power generation. For the progress cost analysis of the construction process, the discrete event simulation method can be used to establish a three-dimensional visualization model of the construction scene, simulate the connection and coordination of key processes such as dam filling, concrete pouring, and electromechanical installation, and optimize the construction organization plan and resource allocation plan. For the impact assessment of the project's ecological environment, the system dynamics method can be used to establish a multi-agent model of the reservoir ecosystem to simulate the impact of the reservoir's water storage on the hydrological regime, species habitat, landscape pattern, etc. in the reservoir area and downstream of the river, and weigh the project benefits and environmental costs.

[0030] After the simulation model is established, the key parameters of the second candidate design scheme are used as input to carry out a large number of simulation tests under different working conditions, generate a large amount of simulation data, and extract effective feedback information from it. Intelligent algorithms such as big data analysis and machine learning are used to classify, cluster, associate and predict the simulation results, evaluate the comprehensive performance of the design scheme, diagnose potential risks and faults, and predict long-term operation trends. The refined feedback information is compared with the original design requirements and constraints to identify the gaps and deficiencies of the design scheme in terms of function, safety, environmental protection, economy, etc., and transform them into operational improvement directions and optimization goals. At the same time, the sensitivity and uncertainty of the design scheme are quantitatively analyzed to determine the design parameters that have a significant impact on the key performance indicators, and carry out parameter optimization in a targeted manner. On this basis, the design optimization model of the previous link is returned to convert the simulation feedback into design knowledge, the weight coefficient in the optimization objective function is corrected, the value range of the optimization variable is adjusted, and the relevant constraints are added or deleted, thereby guiding the optimization search to develop in a direction that is more in line with actual needs. After multiple rounds of iterations, lessons learned from simulation analysis are continuously used to correct the optimization model. At the same time, the optimization results are returned to the simulation test for verification. Ultimately, the target design solution that meets all requirements is obtained through the spiral iteration of simulation, optimization, and correction.

[0031] S202, performing statistical analysis on the hydrological data to obtain hydrological characteristic parameters; Specifically, when conducting statistical analysis on hydrological data, it is necessary to select reasonable statistical methods and models according to the purpose of analysis and data characteristics. Commonly used statistical analysis methods include frequency analysis, cause analysis, correlation analysis, etc. The frequency analysis method obtains design values ​​of different recurrence periods, such as design flood, design low water, design sediment content, etc. by performing frequency statistics on multi-year hydrological sequences, and provides design parameters in flood control, water supply, hydropower generation, etc. Its basic steps include consistency test of hydrological sequences, outliers detection, sequence fitting, parameter estimation, and drawing of frequency curves. The cause analysis method is based on the mechanism of runoff generation and convergence in the basin, analyzes the contribution of factors such as rainfall, temperature, and underlying surface to runoff formation, constructs a "black box" or "gray box" hydrological model, and reveals the inherent laws of the hydrological process. The correlation analysis method examines the correlation and dependence between hydrological elements, and obtains the quantitative relationship between runoff and rainfall, sediment, evaporation, etc., and then conducts medium- and long-term hydrological forecasts. Regardless of the statistical method used, the original hydrological time series needs to be preprocessed, such as outlier removal, missing value interpolation, and stabilization of non-stationary series, to ensure that the data quality meets the analysis requirements.

[0032] After completing the data preprocessing, the corresponding statistical models and algorithms are used to compile and analyze the hydrological data, and a series of hydrological characteristic parameters are extracted from them. Taking frequency analysis as an example, the mean, coefficient of variation, and skewness coefficient of the design flood are obtained through parameter estimation, and the flood frequency curve is drawn to determine the peak flow, flood volume, duration and other parameters of different recurrence periods, and then the formation mechanism and distribution law of the basin flood are analyzed. Taking the cause analysis as an example, the basin water balance equation is established to quantitatively analyze the contribution rate of runoff, confluence, loss and other links to runoff, and the basin underlying surface scenario analysis and runoff change trend prediction are carried out based on this. For the use of reservoirs, it is also necessary to obtain key parameters such as design guarantee rate, benefit storage capacity, flood control storage capacity, and flood regulation calculation results through methods such as wet and dry period analysis and guarantee rate analysis. In the process of parameter extraction, it is necessary to comprehensively and systematically analyze the spatiotemporal differentiation characteristics of the hydrological process, highlight the leading role of key control factors and sensitive parameters, and quantitatively characterize the response relationship between hydrological elements. In expressing the analysis results, intuitive and vivid visual charts should be used, accompanied by easy-to-understand text descriptions, to facilitate use in engineering consultation and solution demonstration.

[0033] S203, establishing a correlation model between terrain characteristic parameters and hydrological characteristic parameters through a curve fitting method, and screening the terrain characteristic parameters and hydrological characteristic parameters based on the correlation model to obtain key characteristic parameters.

[0034] Specifically, when establishing a correlation model between topographic features and hydrological features, it is necessary to select a reasonable curve fitting method, comprehensively considering the nonlinear relationship between the dependent and independent variables, multicollinearity problems, heteroscedasticity problems, etc. Commonly used curve fitting methods include univariate linear regression, multivariate linear regression, nonlinear regression, generalized additive model, etc. Among them, univariate linear regression is suitable for exploring the correlation between a single topographic factor (such as basin area, river network density) and a certain hydrological element (such as peak flow, annual runoff), multivariate linear regression is suitable for comprehensively considering the interactive effects of multiple topographic factors, nonlinear regression is suitable for describing complex nonlinear relationships (such as power function, logarithmic function), and generalized additive model takes into account the linear and nonlinear effects of independent variables, and has higher prediction accuracy. When choosing the model form, it is necessary to consider the model's goodness of fit and prediction effect, as well as the model's simplicity and physical meaning, to avoid overfitting and information redundancy. Generally speaking, the relationship type between variables can be preliminarily determined through scatter plots, residual plots, etc., and then candidate models can be selected based on mechanism analysis and professional knowledge. The performance of different models can be compared through cross-validation, information criteria, etc., and the optimal association model can be finally determined.

[0035] In the process of establishing the correlation model, it is necessary to preprocess the original terrain data and hydrological data. On the one hand, data screening and quality control should be carried out to remove abnormal values ​​and outliers, interpolate missing data, and eliminate inconsistencies; on the other hand, data standardization and normalization should be carried out to eliminate dimension and order of magnitude differences and improve the comparability of variables. For some distributed terrain parameters (such as elevation and slope), it is also necessary to extract their statistical characteristic values ​​(such as mean, variance, quantile), or perform spatial interpolation and basin averaging to match the time scale of hydrological parameters. In the selection of independent variables, it is necessary to fully consider multiple key elements of the basin topography, including basin area, river network density, elevation difference, average slope, curvature, undulation, water system gradient, etc., to reflect the dominant control role of terrain on runoff generation and confluence, and to reduce the correlation between variables as much as possible to improve the stability of the model.

[0036] After constructing the terrain-hydrology correlation model, it is necessary to use appropriate model evaluation indicators, such as determination coefficient, root mean square error, Nash coefficient, etc., to quantitatively judge the model's fitting effect and prediction accuracy. For the correlation model with good performance, its physical mechanism can be further analyzed to explain the response differences of hydrological elements under different terrain conditions and reveal the action mode and sensitivity of key terrain factors. Based on this, a two-dimensional parameter space of "terrain-hydrology" is constructed to screen key characteristic parameters. Specifically, considering factors such as model goodness of fit, parameter significance, and physical significance, the terrain elements that play a decisive role in the hydrological process are identified. For example, the basin area and river network density may have the greatest impact on the peak flow, and the slope and undulation may have the greatest impact on the runoff coefficient. In the parameter screening process, it is necessary to avoid over-simplifying the model, which leads to the loss of key information; it is also necessary to avoid including too many parameters, which leads to excessive model complexity and violates the principle of simplicity. The key characteristic parameters screened can be used to guide the parameter calibration and verification of hydrological models under different terrain conditions, realize the regionalization of model structure and parameters, and improve the applicability and extension of the model.

[0037] S103, determining design constraints according to engineering design requirements, and establishing a three-dimensional design optimization model based on key characteristic parameters as design variables, engineering design requirements, and design constraints; Specifically, after obtaining the engineering design requirements and extracting the key characteristic parameters of terrain data and hydrological data, it is necessary to further determine the design constraints according to the engineering design requirements, and use the key characteristic parameters as design variables. At the same time, a three-dimensional design optimization model is established by combining the engineering design requirements and design constraints. The reason for establishing a three-dimensional design optimization model is that traditional water conservancy engineering design often adopts a two-dimensional plane layout, which makes it difficult to fully consider the influence of terrain and hydrological conditions. The use of three-dimensional modeling can more realistically and accurately depict the complex engineering reality, and optimize the design scheme through digital simulation of the entire process of engineering operation. Incorporating key characteristic parameters into the model as design variables fully reflects the design concept of adapting measures to local conditions, and achieves a balance between engineering benefits and investment by matching engineering layout with site conditions. The constraints determined according to the engineering design requirements limit the value range of the design variables, ensuring the safety, feasibility and economy of the design scheme.

[0038] Specifically, when establishing a three-dimensional design optimization model, first construct a three-dimensional terrain model based on the terrain characteristic parameters and use it as the basic scene for the design. Using digital terrain modeling technology, the terrain surface is divided into TIN (irregular triangulated network) to generate a high-precision terrain surface model. On this basis, the spatial position and layout of major buildings such as the reservoir area, dam site, spillway, diversion tunnel, etc. are determined according to the hydrological characteristic parameters, and a three-dimensional solid model is built. Using the parametric modeling method, the characteristic dimensions, elevation, slope, etc. of the building are used as design variables, and the optimization of the project layout is achieved by adjusting the values ​​of these variables. At the same time, make full use of BIM (Building Information Modeling) technology to integrate the project's attribute information, logical relationships, construction progress, etc. in the three-dimensional model to achieve integrated management of design, construction, and operation and maintenance.

[0039] While constructing the three-dimensional solid model, it is also necessary to establish corresponding constraints and evaluation indicators according to the requirements of engineering design. Design constraints mainly include mandatory constraints such as engineering safety, construction conditions, investment limits, and ecological protection, which limit the values ​​of design variables. For example, for a reservoir dam, it is necessary to determine the verification flood level according to the flood control standard, and use it as a constraint to control the dam crest elevation and spillway size; determine the allowable range of parameters such as the dam foundation depth and slope according to geological conditions; and limit the amount of waste slag and land area during the construction period according to environmental protection requirements. At the same time, according to the functional positioning of the project, it is also necessary to propose evaluation indicators for different aspects such as flood control, water supply, power generation, irrigation, and shipping, and assign weight coefficients to construct a multi-objective evaluation function.

[0040] By taking key characteristic parameters as design variables and establishing a three-dimensional design optimization model in combination with engineering design requirements and constraints, we can fully reflect the systematic view of the interaction and coordinated development between engineering and the environment, and realize multi-disciplinary and multi-objective coupling solutions. Compared with traditional two-dimensional models, three-dimensional models can more comprehensively and dynamically reflect the actual engineering situation, provide a visual and interactive design platform for various disciplines, and improve the efficiency of information sharing and collaborative design. Through parametric modeling technology, the design scheme can be adjusted quickly and flexibly, and its comprehensive benefits can be evaluated in real time, greatly shortening the design cycle. Incorporating terrain and hydrological characteristic parameters into the model can optimize the engineering layout according to local conditions and tailor it to the site, give play to the comparative advantages of the site, and maximize the investment benefits while ensuring safety. In addition, by balancing different goals such as flood control, water supply, and power generation, and reasonably weighing multiple demands such as engineering benefits, ecological impacts, and resettlement of immigrants, we can find the optimal solution that unifies economic, environmental, and social benefits, and achieve sustainable development of engineering construction.

[0041] Based on the above embodiment, as an optional embodiment, a three-dimensional design optimization model is established based on key characteristic parameters as design variables, engineering design requirements and design constraints, including: S301, determining an objective function according to engineering design requirements; Specifically, when establishing a three-dimensional design optimization model, it is first necessary to determine the objective function according to the engineering design requirements. The objective function is a quantitative indicator for judging the quality of the design scheme, reflecting the designer's pursuit and preference for engineering benefits. In actual engineering, the design objectives are often diverse, including both engineering economic benefit objectives (such as minimum engineering investment and maximum engineering benefits) and social ecological benefit objectives (such as optimal flood control safety, highest water supply guarantee rate, and maximum ecological flow). Therefore, it is necessary to weigh and balance various design requirements according to the nature and dominant functions of the project and select an appropriate objective function form. Common objective function forms include linear functions, nonlinear functions, and multi-objective functions. Among them, linear functions are suitable for describing simple linear relationships between design variables and target quantities, such as the linear correlation between construction costs and earthwork volume; nonlinear functions are suitable for describing complex nonlinear effects, such as the nonlinear relationship between power generation benefits and head and flow; multi-objective functions take into account multiple design objectives and normalize them into a single objective function by weighted summation. The weight coefficient reflects the relative importance of different objectives. When selecting the objective function, we must ensure that it can accurately reflect the design intent and simplify the function form as much as possible to avoid unsolvable or difficult-to-converge problems.

[0042] Taking the design of reservoir engineering as an example, if flood control and water supply benefits are the main goals, the following forms of objective functions can be selected: maximizing flood control benefits (such as reducing the number of days when floods exceed the standard and reducing the flooded area), maximizing water supply benefits (such as increasing the water supply guarantee rate and increasing the irrigation area), and taking into account the minimization of project investment (such as reducing the amount of earthwork and optimizing the construction process). The specific form of the objective function can be flexibly set according to the actual project and data support. For example, the flood control target can be expressed as the minimum maximum peak flow during the flood recurrence period, the water supply target can be expressed as the maximum design guarantee rate, and the investment target can be expressed as the minimum total project cost. When multiple design goals are involved, the weight coefficients of each goal can be determined by methods such as hierarchical analysis and expert scoring, and the multi-objective function can be converted into a single objective function, which is convenient for subsequent model solving and scheme comparison. It should be noted that the rationality of the objective function is directly related to the quality of the design results. Therefore, when determining the objective function, it is necessary to pay attention to widely listening to the opinions of all parties, conducting sufficient scheme demonstration, and conducting special research and technical and economic evaluation when necessary to ensure that the objective function can accurately reflect the design needs of the project.

[0043] After clarifying the design objectives, it is necessary to select appropriate decision variables as independent variables of the optimization model. In the three-dimensional design optimization model, decision variables mainly refer to the key characteristic parameters of the basin identified in the early stage, including the controlling indicators of topography, hydrology, engineering, etc., such as basin area, slope, runoff coefficient, reservoir capacity, flood discharge, etc. Through sensitivity analysis and correlation analysis, the key parameters that have a greater impact on the design objectives and are independent of each other can be screened out as decision variables to avoid introducing information redundancy and excessive constraints. When selecting decision variables, it is necessary to fully reflect the main influencing factors of the problem, and to take into account the suitability and convergence of the model, so as to avoid too many variables leading to a sharp increase in the amount of calculation. At the same time, the value range of the decision variables should be reasonably set. On the one hand, it should cover various possible design conditions, and on the other hand, it should conform to engineering practice and experience judgment to avoid unrealistic extreme situations. In addition, certain key parameters (such as reservoir capacity) can be converted into multiple binary variables through discretization processing, and each variable corresponds to an alternative plan, thereby converting the continuous problem into a combinatorial optimization problem and reducing the difficulty of solving.

[0044] S302, constructing a three-dimensional design optimization model based on design variables, design constraints and objective functions.

[0045] Specifically, in the three-dimensional design optimization model, the design variables are the key parameters that affect the change of the objective function. These variables usually include control parameters related to topography, hydrology and the project itself, such as basin area, slope, runoff coefficient, reservoir capacity, flood discharge, etc. When selecting design variables, it is necessary to ensure that the selected variables can fully represent the main influencing factors of the design problem, and to screen out independent parameters that have a greater impact on the objective function through methods such as sensitivity analysis. At the same time, in order to ensure the computational efficiency of the model, it is also necessary to consider avoiding the introduction of too many redundant variables. By reasonably selecting and setting design variables, not only can the solvability of the optimization problem be ensured, but also the efficiency of the solution process can be improved.

[0046] After the design variables and objective function are clear, the third key link of step S302 is to introduce the design constraints into the optimization model. Design constraints usually include limiting factors such as the technical requirements of the project, environmental protection standards, and construction budget. For example, the storage capacity of the reservoir project, the maximum runoff coefficient of the basin, and the minimum flood discharge of the reservoir may all become constraints in the design optimization. These constraints ensure that the optimization results are feasible in the actual project implementation process. By introducing these design constraints into the model, the optimization process can better adapt to the actual environment and avoid generating design solutions that do not meet actual requirements.

[0047] Finally, by integrating the objective function, design variables and design constraints, a three-dimensional design optimization model can be constructed. At this point, the construction of the model must not only meet the needs of the design objectives, but also take into account various engineering constraints to ensure that the model can accurately describe the overall picture of the design problem. In this process, the use of appropriate optimization algorithms can perform global searches in complex design spaces to find the optimal design solution. Through this step, complex design problems with multiple objectives and multiple constraints can be effectively transformed into a systematic and solvable optimization problem, promoting the development of water conservancy project design in an efficient and scientific direction.

[0048] S104, solving the three-dimensional design optimization model by using a genetic algorithm to obtain a set of non-inferior solutions as the first candidate design solutions; Specifically, after establishing a three-dimensional design optimization model containing key characteristic parameters, design constraints, evaluation indicators and other elements, it is necessary to select a suitable optimization algorithm to solve the model and obtain a design solution that satisfies the constraints and achieves non-inferior optimality on the objective function. Considering that water conservancy project design optimization problems often have the characteristics of many variables, strong constraints, and complex objectives, traditional mathematical programming methods are difficult to work, and intelligent optimization algorithms, especially genetic algorithms, have shown unique advantages in solving such problems. Therefore, the three-dimensional design optimization model is solved by genetic algorithms to obtain a set of non-inferior solutions as the first candidate design solution. The reason for using genetic algorithms is that they have the characteristics of strong global optimization ability, good robustness, and strong versatility. They can quickly converge to the global optimal solution in a complex multi-peak search space, and provide multiple candidate solutions that meet Pareto optimality, which is convenient for decision makers to weigh and choose.

[0049] Specifically, when applying genetic algorithms to solve three-dimensional design optimization models, it is first necessary to determine the chromosome encoding method, map the design variables to the gene positions on the chromosome, and define the fitness function, selection operator, crossover operator and mutation operator. Real number encoding is usually used to divide the value range of each design variable into several discrete intervals, and a real number is used to represent the serial number of the interval, thereby forming a chromosome composed of multiple design variables. The fitness function is constructed according to the design goals and evaluation indicators, and the quality of chromosomes is quantitatively evaluated to guide the evolutionary direction of the population. The selection operator selects a certain number of individuals from the parent generation according to the fitness value of the individual, and inherits them to the next generation, ensuring the inheritance of excellent genes and the accumulation of advantages. The crossover operator generates new offspring chromosomes by exchanging and recombining the gene fragments of two parent chromosomes, realizing the expansion of the search space and the combination innovation of optimization variables. The mutation operator randomly perturbs certain gene positions on the chromosome with a certain probability, jumps out of the local optimum, and increases the diversity of the population.

[0050] After determining the parameters of the genetic algorithm, the initial population is iterated and evolved. Generally, the initial population is generated by random method, that is, a certain number of initial solutions are formed by randomly selecting combinations within the range of the design variables. Then the fitness of each individual is evaluated, the individuals with low fitness are eliminated, and the individuals with high fitness are retained. The retained individuals are crossover and mutation operations are performed according to the preset crossover probability and mutation probability to generate new offspring individuals, which are added to the population to form a new generation. This iteration is carried out until the preset evolutionary generation is reached or the convergence condition is met. During the evolution process, the diversity and convergence of the population are maintained through methods such as non-dominated sorting and crowding calculation, and the Pareto optimal solution set is continuously updated and output as the first candidate design solution.

[0051] Based on the above embodiment, as an optional embodiment, the three-dimensional design optimization model is solved by a genetic algorithm to obtain a set of non-inferior solutions as the first candidate design solution, including: S401, randomly generating an initial population including a plurality of individuals in a design space of a three-dimensional design optimization model, wherein each individual corresponds to a set of value combinations of design variables; Specifically, during the implementation process, it is first necessary to clarify the design space of the constructed three-dimensional design optimization model. This space consists of multiple design variables and corresponding value ranges. In order to effectively search this design space, multiple individuals are first randomly generated as the initial population. Each individual represents a set of value combinations of design variables, which usually include key parameters such as reservoir capacity, basin area, slope, and flood discharge. Through random generation, it is ensured that the initial population covers a variety of possible solutions in the design space, providing a variety of initial solutions for the subsequent optimization process. This random generation method can avoid falling into the local optimal solution too early and provide a wider search space for subsequent algorithms, thereby increasing the probability of finding the global optimal solution.

[0052] Specifically, when the random algorithm is used to generate the initial population, possible intervals are first delineated according to the range of values ​​of each design variable, and then a set of values ​​are randomly selected within these intervals as the design variable values ​​of each individual. Multiple such individuals are generated so that the entire initial population can cover different parts of the design space. The purpose of this is to increase the breadth of the search space during the evolution of the genetic algorithm through diversified initial solutions, providing greater possibilities for subsequent optimization.

[0053] S402, calculating the fitness value of each individual in the initial population according to the objective function in the three-dimensional design optimization model, and sorting the individuals according to the fitness value to obtain sorted individuals; Specifically, in order to evaluate the quality of each individual in the initial population, each individual needs to be substituted into the objective function for calculation to obtain the fitness value of the individual. The fitness value is a quantitative indicator of individual performance, usually expressed as a numerical value. The higher the value, the better the design scheme of the individual. In multi-objective optimization problems, the fitness value is usually a comprehensive evaluation of the values ​​of each objective function, and a total fitness value can be obtained by weighted summation or other methods. In this way, by calculating the fitness value, the performance of each individual on the objective function can be accurately quantified. Next, based on the calculated fitness value, all individuals are sorted in order from high to low. The sorted individuals reflect the design scheme that performs best in the current iteration, as well as the design scheme that performs poorly. The purpose of this is to provide a basis for subsequent selection operations. Individuals with high fitness are retained first, so that excellent design schemes can be passed on to the next generation, thereby promoting the overall population to develop towards a better solution set.

[0054] S403, screening based on the ranked individuals, selecting the top N individuals with the highest ranking into the parent individual group, and randomly selecting two parent individuals from the parent individual group for crossover operation to obtain offspring individuals; Specifically, the screening process needs to select the first N individuals with the highest fitness values ​​according to the fitness values ​​after sorting. For example, if the first N individuals in the sorted individuals represent different design solutions, these N individuals will be selected into the parent individual group. At this time, the parent individual group contains the best solution in the current design space, providing high-quality seeds for subsequent genetic operations. Next, a crossover operation is performed within the parent individual group. The crossover operation is a key process in the genetic algorithm, and its purpose is to generate new offspring individuals by combining the gene fragments of the parent individuals. Specifically, two individuals are randomly selected from the parent individual group as parents for the crossover operation. The crossover operation combines the excellent characteristics of the two parents by exchanging the parent gene fragments to generate a new combination of design variables, thereby providing more diverse solutions for the optimization of the next generation. As a result of the crossover operation, the generated offspring individuals may combine the advantages of the two parent individuals and have better fitness, which may further improve the optimization effect in subsequent iterations.

[0055] S404, merging the parent individuals and the offspring individuals to form a new population, and performing evolution processing on the new population until the latest population evolution reaches a preset convergence condition to obtain a convergent population; Specifically, during the merging process, the parent individuals and the offspring individuals are usually integrated according to the fitness value or other preset criteria. You can choose to merge all the parent individuals and offspring individuals, or select only some individuals with higher fitness to join the new population according to certain rules. Regardless of which method is chosen, the merged new population must be guaranteed to have sufficient diversity so that the algorithm can continue to explore the design space and avoid falling into the local optimal solution.

[0056] Next, the new population is subjected to evolutionary processing. The core goal of evolutionary processing is to gradually improve the fitness of individuals in the population through multiple iterations, and promote the overall population to converge to the global optimal solution. During the evolutionary process, the genetic algorithm continuously optimizes the population through operations such as crossover and mutation, and promotes the improvement of the design solution by selecting excellent individuals and introducing new gene mutations. The specific evolutionary process includes multiple fitness evaluations, individual screening, crossover and mutation operations on the new population until the fitness value of the population converges to a preset standard.

[0057] In this process, the preset convergence condition can be reaching a certain number of iterations, or the change in population fitness is less than a certain threshold. The convergence standard may be that the fitness change of the best individual in the population tends to be stable, or that the fitness value of the population reaches a certain target value. Only when the convergence condition is met will the genetic algorithm stop evolving and obtain the final converged population.

[0058] S405, extracting a group of non-inferior solutions from the converged population, and taking the group of non-inferior solutions as the first candidate design solutions.

[0059] Specifically, a non-inferior solution refers to a solution that is not worse than other individuals in some objective functions and cannot be surpassed by other individuals in other objective functions. For multi-objective optimization problems, the Pareto dominance criterion is often used to determine non-inferior solutions, that is, when a solution is not inferior to another solution in all objective functions and is better in at least one objective, the former dominates the latter. Therefore, in this step, it is necessary to extract all non-inferior solutions from the convergent population through the Pareto sorting method and ensure that these solutions have no other solutions in the objective space that can dominate them.

[0060] The extraction process can be completed by the non-dominated sorting algorithm. First, the objective function value of each individual is compared to find those individuals that cannot be dominated by any other solution, and these individuals are regarded as non-inferior solutions. The reason for this is that non-inferior solutions represent a set of design solutions in the design space that perform well under different objectives. They are the Pareto frontier solution set in multi-objective optimization problems. By extracting these non-inferior solutions, multiple possible high-quality design solutions can be provided for subsequent decisions, allowing decision makers to make choices under different design requirements and trade-offs.

[0061] S105, optimizing the first candidate design solution by using a preset reinforcement learning model to obtain a second candidate design solution; Specifically, the first candidate design scheme obtained by solving the three-dimensional design optimization model through genetic algorithm has achieved the non-inferior optimal solution of multiple objectives under static conditions. However, in the dynamic evolution of engineering design, the actual effect of the design scheme is still affected by many uncertain factors, and there may be new conflicts and contradictions between different objectives and between short-term and long-term benefits. In order to further improve the robustness, coordination and sustainability of the design scheme, it is necessary to re-optimize the first candidate design scheme in a dynamic environment. Here, a preset reinforcement learning model is adopted. Through the continuous interaction between the intelligent subject and the environment, the design scheme is improved by trial and error under the goal of maximizing long-term benefits, and the second candidate design scheme after dynamic optimization is obtained. The reinforcement learning model is introduced because it can learn and adapt in a complex and changing environment, optimize the entire decision sequence through delayed feedback learning of design decisions, weigh various pros and cons in strategy improvement, and seek the best balance in exploration and utilization, so as to obtain a robust optimal solution that is stably dominated in the dynamic game.

[0062] Specifically, when applying the reinforcement learning model to optimize the first candidate design scheme, it is first necessary to express the water conservancy project design decision problem as a Markov decision process, and clarify the state space, action space and reward function. The state space contains all environmental variables that affect the design scheme, such as random processes such as runoff, sediment, geology, and meteorology, the state attributes of each building during the operation of the project, and the ecological environment of the reservoir area. The action space is composed of various design parameters and their adjustment methods, such as dam type, dam height, dam crest width, spillway layout, construction diversion scheme, etc., and specific design decisions are generated by optimizing these parameters. The reward function measures the immediate benefits of a decision in a certain state, and needs to comprehensively consider different benefits such as flood control, water supply, power generation, ecology, as well as construction costs, immigration losses, environmental costs, etc. The Markov property requires that the next state is only related to the current state and action, and has nothing to do with the past state and action, which can be met by reasonably designing state variables.

[0063] After constructing the Markov decision model, reinforcement learning algorithms such as Q learning are used to establish a mapping strategy from state to action. This strategy function is updated through continuous trial and error learning, so that it can cope with various possible state combinations and make the global optimal design decision sequence. Q learning uses the value function approximation method to use Q value to represent the long-term expected return of taking a certain action in a certain state, and uses the Bellman equation to iteratively update the Q value. In the process of time difference learning, experience is continuously accumulated and decision-making is improved. In the early stage of learning, in order to explore more state-action combinations within a limited training time, the ε-greedy strategy is usually adopted, that is, actions are randomly selected with a greater probability, and the action with the largest current Q value is selected with a certain probability. As learning deepens, the probability of random exploration is continuously reduced, and the use of existing strategies is increased. At the same time, mechanisms such as experience replay are used to randomly extract historical state-action-return data for repeated learning, accelerate convergence, and improve stability. In the learning process, in order to balance short-term and long-term returns, a discount factor is introduced to attenuate future returns, and at the same time, negative returns are set to punish behaviors that violate constraints. After learning reaches convergence, the optimal design strategy is applied to the first candidate design solution to generate a dynamically optimized second candidate design solution.

[0064] By optimizing the first candidate design scheme through the reinforcement learning model, we can seek a robust optimal solution in the dynamic game between engineering design and the external environment, and improve the adaptability of the design scheme to uncertainty. Compared with static optimization, reinforcement learning can consider the environmental changes and benefit dynamics of the entire life cycle of the project operation in design optimization, and continuously correct the optimization direction through rolling improvements to design decisions, so as to obtain the optimal strategy with continuity and consistency on a time scale. Through repeated attempts at different states and decisions, reinforcement learning has learned a lot of successful experiences and failed lessons, and can predict the comprehensive benefits of the design scheme under various possible situations, and make optimization trade-offs to maximize the overall benefits under the premise of satisfying constraints. Taking the genetically optimized scheme as the starting point of reinforcement learning, through the collaborative optimization of two intelligent algorithms, the advantages of the two can be integrated to make up for the shortcomings of a single method, while improving the convergence speed and broadening the optimization vision, further improving the quality and effect of the design scheme. At the same time, the series of alternative optimization schemes provided by reinforcement learning, through interactive games with different environments and goals, present different utility attributes and adaptation characteristics, further enriching the choice space of decision makers.

[0065] Based on the above embodiment, as an optional embodiment, the first candidate design solution is optimized by a preset reinforcement learning model to obtain a second candidate design solution, including: The first candidate design scheme is taken as the starting point of the preset reinforcement learning model, and the engineering design requirement is taken as the action space of the preset reinforcement learning model. The first candidate design scheme is optimized to obtain the second candidate design scheme.

[0066] Specifically, the design space of the reinforcement learning model is defined as engineering design requirements, which include multiple objectives such as flood control benefits, water supply capacity, and construction costs. These objectives will serve as the action space of the reinforcement learning model, and each action represents a change or adjustment to the design plan. For example, for key design variables such as watershed area and reservoir capacity, the reinforcement learning model interacts with the environment to explore different adjustment methods in order to optimize the objective function and achieve the design goal.

[0067] The reinforcement learning model optimizes the first candidate design through a mechanism based on rewards and penalties. First, the reinforcement learning model evaluates the matching degree of the current design, that is, the actual performance of the first candidate design, with the design requirements, and then calculates the reward signal according to the design goal. The reward signal guides the model to adjust the design variables so that the design is closer to the ideal goal in the next step. When the adjustment operation selected by the model can improve the design on the objective function, it will receive a positive reward; on the contrary, if the adjustment causes the objective function to perform worse, it will be punished. Through multiple rounds of training, the reinforcement learning model can gradually learn which adjustments are most likely to improve the quality of the design, thereby gradually optimizing the first candidate design.

[0068] This process will eventually generate a second candidate design, which is a new solution that has been optimized and adjusted many times under the guidance of the reinforcement learning model. Compared with the first candidate design, the second candidate design performs better in the objective function and meets more detailed and optimized design requirements. In this way, the reinforcement learning model can not only further improve the initial results produced by the genetic algorithm, but also explore more potential optimization space to ensure that the design more accurately meets the actual needs.

[0069] S106, performing simulation analysis based on the second candidate design solution to obtain feedback information, and optimizing the second candidate design solution based on the feedback information to obtain a target design solution.

[0070] Specifically, when conducting simulation analysis of the second candidate design scheme, firstly, a digital twin simulation system for the entire life cycle is built according to the content and requirements of the engineering design. Using digital technologies such as BIM, GIS, and IOT, a high-precision three-dimensional solid model integrating all-factor information such as design parameters, topography and geology, construction technology, operating conditions, and management strategies is constructed, and then coupled with models of professional fields such as hydrodynamics, structural dynamics, and ecological environment to form a multi-physical field simulation system that reflects the entire life cycle of the water conservancy project. On this basis, the various parameters of the second candidate design scheme are input into the simulation system to simulate various possible conditions during the construction and operation stages of the project, and to evaluate the applicability and reliability of the design scheme. Using technical means such as numerical simulation, physical model testing, and virtual reality, the key issues of the building's stress and strain, seepage stability, seismic performance, siltation, and water environment changes are comprehensively examined, with a focus on the weak links and accident conditions of the design scheme. At the same time, full use is made of the monitoring data for synchronous verification and dynamic correction of the simulation model to improve the accuracy and credibility of the simulation analysis.

[0071] Through systematic and comprehensive simulation analysis, we can obtain an objective and quantitative evaluation of the second candidate design, gain insight into its internal mechanism and evolution law, predict the risks and challenges it faces, and form feedback information from the whole process, multiple angles, and in depth. On this basis, we follow the principle of promoting optimization through feedback and guiding improvement through evaluation, focus on the weak links and key constraints revealed by simulation analysis, take targeted optimization measures, and revise and improve the second candidate design again. In view of the safety and stability of hydraulic structures, we optimize the design parameters such as dam type, dam site, dam body material, foundation treatment, and construction diversion to enhance the integrity, durability, and disaster resistance of the structure; in view of the movement of water and sand and siltation, we optimize the reservoir capacity and type, the layout of inlets and outlets, and the strategies for water and sand regulation to improve the distribution of water flow fields in the reservoir area, control siltation, and extend the service life of the reservoir; in view of water environment protection and ecological impact issues, we optimize the construction layout, the site selection of waste dumps, artificial fishways, ecological flow, and vegetation restoration to minimize engineering interference and protect and restore the river ecosystem. After multiple rounds of iterative optimization, while ensuring the design effect and constraints, we further explore the optimization potential of the design scheme, weigh various influencing factors, and balance multiple target requirements, and finally obtain a comprehensive, balanced, dynamically adaptable, efficient and feasible target design scheme.

[0072] Based on the above embodiment, as an optional embodiment, simulation analysis is performed based on the second candidate design solution to obtain feedback information, including: S501, converting each design solution in the second candidate design solution into a parameterized three-dimensional solid model; Specifically, the design variables contained in each design scheme, such as reservoir capacity, dam height, flood discharge, etc., will be connected to the 3D modeling software in a parametric way. Through these design variables, the various geometric features of the 3D model can be defined and the corresponding solid model can be generated. For example, the reservoir capacity determines the size and shape of the dam, and the slope of the basin affects the path of the water flow and the water storage capacity. These key variables will be dynamically adjusted through the parametric input in the modeling software to generate a complete 3D model.

[0073] In this way, the various elements of the design are transformed into operational and computable solid models, which can provide an intuitive and accurate model basis for subsequent simulation analysis. Such a three-dimensional solid model is not only highly accurate, but can also be further calculated and tested in actual engineering. For example, through simulation analysis such as fluid dynamics simulation and earthwork volume estimation, the model can show the specific impact of different design variables on engineering results.

[0074] S502, meshing the three-dimensional solid model to obtain a computational mesh, and performing simulation analysis based on the computational mesh to obtain feedback information.

[0075] Specifically, the meshing process first selects the appropriate mesh type and size based on the geometry and physical properties of the three-dimensional solid model. For water conservancy project design, such as reservoirs, dams and other structures, structured meshes or unstructured meshes are usually used to more accurately simulate the physical phenomena in different areas. For more complex geometries, such as the inside of the dam body and spillways, the density of meshing should be appropriately increased to ensure the accuracy of the calculation. During the division process, special attention should also be paid to some key areas, such as areas where the water velocity changes greatly or areas with higher pressure. The mesh should be encrypted in these areas to improve the accuracy of the simulation results.

[0076] After the grid division is completed, simulation analysis is performed based on the obtained calculation grid. The content of the simulation analysis is mainly based on the physical models of water flow, earthwork, stress, etc. For example, the flow velocity and flow change of the water flow, the stress condition of the dam structure, etc. Through numerical calculation, simulation analysis can provide specific feedback information for the design scheme, such as the distribution of water flow, the stability of the structure, possible risk points, etc. This feedback information reflects the performance of the design scheme under actual working conditions, which helps to discover potential deficiencies and optimization space.

[0077] See also Figure 2 , Figure 2 The present invention provides a three-dimensional water conservancy project design system architecture diagram, which may include: Data acquisition module 1, used to obtain topographic data, hydrological data and engineering design requirements of the project area; Feature extraction module 2, used to extract features from terrain data and hydrological data to obtain key feature parameters; Model building module 3, used to determine design constraints according to engineering design requirements, and to establish a three-dimensional design optimization model based on key characteristic parameters as design variables, engineering design requirements and design constraints; A solution module 4 is used to solve the three-dimensional design optimization model by using a genetic algorithm to obtain a set of non-inferior solutions as the first candidate design solution; A first optimization module 5, configured to optimize the first candidate design solution by using a preset reinforcement learning model to obtain a second candidate design solution; The second optimization module 6 is used to perform simulation analysis based on the second candidate design solution to obtain feedback information, and optimize the second candidate design solution based on the feedback information to obtain a target design solution.

[0078] It should be noted that: when the system provided in the above embodiment realizes its functions, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0079] Please refer to Figure 3 The application also discloses an electronic device. Figure 3 The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, and at least one communication bus 302 or end-to-end wireless communication.

[0080] The communication bus 302 is used to realize the connection and communication between these components.

[0081] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0082] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0083] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field~Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 301 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented separately through a chip.

[0084] Among them, the memory 305 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read~Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally be at least one storage system located away from the aforementioned processor 301. Refer to Figure 3 , the memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a three-dimensional design method for a water conservancy project.

[0085] exist Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call the application program storing the three-dimensional design method of the water conservancy project in the memory 305, and when executed by one or more processors 301, the electronic device 300 executes one or more methods in the above-mentioned embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited to the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application. In the above embodiments, the description of each embodiment has its own emphasis, and the part that is not described in detail in a certain embodiment can refer to the relevant description of other embodiments.

[0086] In the several implementations provided in this application, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are only schematic, such as the division of modules, which is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, indirect coupling or communication connection of systems or modules, which can be electrical or other forms.

[0087] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0088] The present application also provides a computer storage medium that can store multiple instructions, which are suitable for being loaded and executed by a processor as described above. Figure 1 The three-dimensional design method of the water conservancy project in the embodiment shown in the figure can be found in the specific implementation process. Figure 1 The specific description of the illustrated embodiment will not be repeated here.

[0089] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules.

[0090] If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0091] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0092] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A three-dimensional design method for water conservancy projects, characterized in that: The method comprises: Obtain topographic data, hydrological data and engineering design requirements for the project area; Extracting features from the terrain data and hydrological data to obtain key feature parameters; Determining design constraints according to the engineering design requirements, and establishing a three-dimensional design optimization model based on the key characteristic parameters as design variables, the engineering design requirements and the design constraints; Solving the three-dimensional design optimization model by using a genetic algorithm to obtain a set of non-inferior solutions as a first candidate design solution; Optimizing the first candidate design solution by using a preset reinforcement learning model to obtain a second candidate design solution; A simulation analysis is performed based on the second candidate design solution to obtain feedback information, and the second candidate design solution is optimized based on the feedback information to obtain a target design solution.

2. The method according to claim 1, characterized in that: The feature extraction of the terrain data and the hydrological data to obtain key feature parameters includes: Converting the terrain data into a terrain digital elevation model, and performing feature extraction based on the terrain digital elevation model to obtain terrain feature parameters; Performing statistical analysis on the hydrological data to obtain hydrological characteristic parameters; A correlation model between the terrain characteristic parameters and the hydrological characteristic parameters is established by a curve fitting method, and the terrain characteristic parameters and the hydrological characteristic parameters are screened based on the correlation model to obtain the key characteristic parameters.

3. The method according to claim 1, characterized in that The establishing of a three-dimensional design optimization model based on the key characteristic parameters as design variables, the engineering design requirements and the design constraints includes: Determining an objective function according to the engineering design requirements; The three-dimensional design optimization model is constructed based on the design variables, the design constraints and the objective function.

4. The method according to claim 1, characterized in that The three-dimensional design optimization model is solved by a genetic algorithm to obtain a set of non-inferior solutions as a first candidate design solution, including: In the design space of the three-dimensional design optimization model, an initial population including a plurality of individuals is randomly generated, wherein each individual corresponds to a set of value combinations of design variables; According to the objective function in the three-dimensional design optimization model, the fitness value of each individual in the initial population is calculated, and the individuals are sorted according to the fitness value to obtain sorted individuals; Screening is performed based on the ranked individuals, selecting the top N individuals with the highest ranking into the parent individual group, and randomly selecting two parent individuals from the parent individual group for crossover operation to obtain offspring individuals; The parent individuals and the offspring individuals are combined to form a new population, and the new population is subjected to evolution processing until the latest population evolution reaches a preset convergence condition to obtain a convergent population; A group of non-inferior solutions is extracted from the converged population, and the group of non-inferior solutions is used as the first candidate design scheme.

5. The method according to claim 1, characterized in that The step of optimizing the first candidate design solution by using a preset reinforcement learning model to obtain a second candidate design solution includes: The first candidate design scheme is taken as the starting point of a preset reinforcement learning model, and the engineering design requirement is taken as the action space of the preset reinforcement learning model. The first candidate design scheme is optimized to obtain a second candidate design scheme.

6. The method according to claim 1, characterized in that The performing simulation analysis based on the second candidate design solution to obtain feedback information includes: Converting each of the second candidate design solutions into a parameterized three-dimensional solid model; The three-dimensional solid model is meshed to obtain a computational mesh, and simulation analysis is performed based on the computational mesh to obtain the feedback information.

7. The method according to claim 1, characterized in that The optimizing the second candidate design scheme based on the feedback information to obtain a target design scheme includes: Analyze the feedback information to determine defect indicators; Determine a design parameter associated with the defect indicator, and optimize the second candidate design scheme based on the design parameter to obtain the target design scheme.

8. A three-dimensional design system for water conservancy projects, characterized in that: The system comprises: Data acquisition module, used to obtain topographic data, hydrological data and engineering design requirements of the project area; A feature extraction module is used to extract features from the terrain data and hydrological data to obtain key feature parameters; A model building module, used to determine design constraints according to the engineering design requirements, and to build a three-dimensional design optimization model based on the key characteristic parameters as design variables, the engineering design requirements and the design constraints; A solution module, used for solving the three-dimensional design optimization model by using a genetic algorithm to obtain a set of non-inferior solutions as a first candidate design solution; A first optimization module, used to optimize the first candidate design solution by using a preset reinforcement learning model to obtain a second candidate design solution; The second optimization module is used to perform simulation analysis based on the second candidate design solution to obtain feedback information, and optimize the second candidate design solution based on the feedback information to obtain a target design solution.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It includes a processor, a memory and a transceiver, the memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method according to any one of claims 1 to 7.

Images