Parameter optimization method for digital model generative design

Abstract

The invention discloses a parameter optimization method for digital model generative design, and compared with a traditional engineering parameter optimization method, such as manual adjustment based on experience or single algorithm optimization, the parameter optimization process for the digital model generative design introduces a multi-level modeling and multi-objective optimization mechanism. According to the method, the scientificity and systematicness of a design scheme are remarkably improved, through a full-link process from initial model calibration to global and local modeling and then to multi-target tradeoff and result visualization, computing resources and design intelligence are effectively integrated, an intelligent optimization path of'data driving + experience compensation 'is achieved, and the method is suitable for large-scale popularization and application. According to the method, the design efficiency, the model precision and the scheme quality are improved, the design risk and the trial and error cost are greatly reduced, through systematic calibration and formulation expression of the initial parameters, the digital model can be in a reasonable engineering response interval at the initial stage of optimization, and invalid iteration caused by initial modeling deviation is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of hydropower engineering, and in particular to a parameter optimization method for digital model generative design. Background Art

[0002] With the widespread adoption of technologies such as BIM (Building Information Modeling), digital twins, and high-performance computing, digital models have become essential tools for the design and analysis of complex infrastructure. In dam projects, due to the high demands placed on structural safety, functional adaptability, and economic efficiency, the efficient and precise optimization of design parameters while meeting multiple constraints has become a key issue in current engineering design.

[0003] The current digital design process for dam structures still suffers from widespread problems such as unstable parameter optimization results, frequent local convergence traps, and insufficient model response accuracy. Existing optimization methods often perform unidirectional or static analysis of stability indicators and failure boundary parameters, lacking a dynamic modeling mechanism for the nonlinear coupling relationship between the two. This results in the model's inability to accurately identify potential defect areas, impacting the effectiveness of structural safety assessments. Furthermore, traditional methods lack robust feedback control mechanisms and dynamic positioning assessment indicators, making it difficult for designers to adjust optimization directions in real time during the iterative process, impacting design efficiency and the reliability of optimization results. Summary of the Invention

[0004] The main purpose of this application is to provide a parameter optimization method for generative design of digital models to solve the problems raised by the above background technology.

[0005] To achieve the above objectives, this application provides the following technical solutions: A parameter optimization method for generative design of digital models, characterized by the following specific steps: S1. Initial model calibration: calibrate the basic digital model according to the design requirements and constraints of the project; S2. Global model establishment: conducting a global search, combining physical constraints and functional requirements, and conducting a multi-dimensional evaluation of the dam design scheme to generate a global model based on the basic digital model; S3. Local model establishment: using genetic algorithm combined with local search strategy, local model is established based on the global model, and the local model is analyzed and tested; S4, multi-objective optimization, uses multi-objective optimization algorithm to optimize and reanalyze the local model; S5. Feedback: Feedback each optimization parameter to the visualization terminal.

[0006] Preferably, in step S1, the specific steps of calibrating the basic digital model are as follows: S1.1. Extract key design parameters based on the project data, specification requirements, and environmental conditions to generate structural load data sets, strength coordination data sets, and load-settlement parameter data sets. Standardize the generated data to ensure consistency between the data structure and the modeling interface. The structural load data group includes the dead load value YZ and the live load value RZ; The strength match data set includes the water-binder ratio value SJB and the cubic compressive strength value LFK; The bearing settlement parameter data group includes the foundation bearing capacity characteristic value DJT and the foundation deformation modulus value DJM; S1.2. Use digital modeling tools to construct a preliminary digital structural model. Input the processed structural load data set, strength coordination data set, and load-settlement parameter data set into the digital model to construct the initial calibration value JCZ. The specific calculation formula is as follows:

[0007] Where: YZ is the dead load value, RZ is the live load value, LFK is the cubic compressive strength value, SJB is the water-cement ratio value, DJT is the foundation bearing capacity characteristic value, and DJM is the foundation deformation modulus value.

[0008] Preferably, in step S2, the specific steps of establishing the global model are as follows: Step S2.1: Collect the design goals for the safety, economy, and functional integrity of the dam project, construct a functional requirement data set, and construct a physical constraint data set based on the project specifications and operating conditions. Standardize the generated data to ensure that the data structure is consistent with the modeling interface. The functional requirement data group includes the effective storage capacity value KRZ, the downstream flow regulation capacity value XHZ and the structural stiffness value GTZ; The physical constraint data set includes the anti-sliding stability safety factor AQZ, the maximum allowable displacement value of the dam body WYZ, and the maximum compressive stress value of the dam foundation YLZ; Step S2.2: Based on the established multi-objective optimization model, genetic algorithm and particle swarm optimization are used to iterate the initial calibration value JCZ according to the functional requirement data set and the physical constraint data set in the multi-dimensional parameter space to generate the global calibration value QJY. The specific calculation formula is as follows:

[0009] Where: KRZ is the effective storage capacity, XHZ is the discharge flow regulation capacity, GTZ is the structural stiffness, AQZ is the anti-sliding stability safety factor, WYZ is the maximum allowable displacement of the dam body, YLZ is the maximum compressive stress of the dam foundation, and JCZ is the initial calibration value.

[0010] Preferably, in step S3, the specific steps of establishing the local model are as follows: S3.1. Collect stability and failure boundary parameters, construct optimization solution verification data sets, stability data sets, and failure boundary data sets, and standardize the generated data to ensure that the data structure is consistent with the modeling interface; The optimization solution verification data set includes deformation BXL and deformation stiffness BXG; The stability data set includes anti-slip force KHL and sliding force HDL; The failure boundary data set includes fracture toughness DLR and ultimate load JHZ; S3.2. By coupling the stability data set and the failure boundary data set with the optimization solution verification data set and the global calibration value QJY, respectively, a genetic algorithm is used to perform local iterative updates on the stability and failure boundaries to synchronously adjust multiple data sets, thereby generating a stability optimization reference coefficient WDX and a failure boundary optimization reference coefficient PBJ; S3.3. Analyze the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ to determine the local optimization range.

[0011] Preferably, in step S3.2, the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ are specifically calculated as follows:

[0012]

[0013] Where: BXL is the deformation, BXG is the deformation stiffness, KHL is the anti-sliding force, HDL is the sliding force, DLR is the fracture toughness, JHZ is the ultimate load, and QJY is the global calibration value; a1 and a2 are empirical adjustment factors. The value range of a1 is [0.5, 2], and the value range of a2 is [0.5, 3]. The specific values ​​of a1 and a2 are adjusted by the user. b1 and b2 are nonlinear exponents. The value range of b1 is [0.3, 1.5], and the value range of b2 is [0.5, 2.5]. The specific values ​​of b1 and b2 are adjusted and set by the user.

[0014] Preferably, in step S3.3, the specific analysis of the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ is as follows: when When , it means that there is no abnormality in the digital model in terms of structural positioning and no optimization is required; when When , it means that the digital model has an abnormality in structural positioning and needs to be optimized; when When , it means that there is no abnormality in the digital model on the damage boundary and no optimization is required; when , it means that the digital model is abnormal on the failure boundary and needs to be optimized.

[0015] Preferably, in step S4, the specific steps of multi-objective model optimization are as follows: S4.1. Collect supplementary parameters, construct a supplementary data set for structural stability and an extended data set for failure behavior, and standardize the generated data to ensure that the data structure is consistent with the modeling interface; The stability supplementary data set includes support spacing ZCJ, lateral displacement CYW and horizontal equivalent stiffness SPG; The extended data set of failure behavior includes crack length LFC, peak stress FYL, and crack tip radius LJB; S4.2. Use a genetic algorithm to perform local iterative updates on the structural stability supplementary data set and the failure behavior extended parameter set, respectively, with the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ to synchronously adjust multiple data. The structural stability supplementary data set and the failure behavior extended parameter set are coupled with the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ to generate the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB. S4.3. Analyze the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB to locate the defect area of ​​the model.

[0016] Preferably, in step S4.2, the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB are calculated by the following formulas:

[0017]

[0018] Where: ZCJ is the support spacing, CYW is the lateral displacement, SPG is the horizontal equivalent stiffness, LFC is the crack length, FYL is the peak stress, LJB is the crack tip radius, WDX is the stability optimization reference coefficient, and PBJ is the failure boundary optimization reference coefficient.

[0019] Preferably, in step S4.3, the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB are specifically analyzed as follows: when When , it means that the accuracy of digital model generation in structural positioning is at the first level and no adjustment is required; when When , it means that the accuracy of digital model generation in structural positioning is at the second level, and the structural positioning model needs to be locally optimized; when When , it means that the accuracy of digital model generation in structural positioning is at level 3, and the structural positioning model needs to be fully optimized; when When , it means that the accuracy of digital model generation on the damage boundary is at the first level and no adjustment is required; when , which means that the accuracy of digital model generation on the damage boundary is in the secondary state, and the damage boundary model needs to be locally optimized; when When , it means that the accuracy of digital model generation on the damage boundary is at level 3, and the damage boundary model needs to be fully optimized.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. Compared to traditional engineering parameter optimization methods, such as manual adjustments based on experience or single-algorithm optimization, this parameter optimization process for generative digital model design introduces a multi-level modeling and multi-objective optimization mechanism, significantly improving the scientific and systematic nature of the design solution. Through a comprehensive process from initial model calibration to global and local modeling, and then to multi-objective trade-offs and result visualization, this method effectively integrates computing resources and design intelligence, achieving a "data-driven + experience-based" intelligent optimization path. This not only improves design efficiency, model accuracy, and solution quality, but also significantly reduces design risk and trial-and-error costs.

[0021] 2. Through the systematic calibration and formulaic expression of the initial parameters, the digital model can be in a more reasonable engineering response range at the initial stage of optimization, reducing invalid iterations caused by initial modeling deviations, effectively improving optimization efficiency and reducing the waste of computing resources. The design variables involved in the initial calibration value JCZ indicator can serve as an important reference for the construction of the objective function and the evaluation of parameter sensitivity in the subsequent optimization process. It has clear physical meaning and engineering relevance, and provides data support for the convergence and robustness of the multi-objective optimization algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a step diagram of the application method. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0024] The terms "first," "second," and "third" in this application are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, features specified as "first," "second," or "third" may explicitly or implicitly include at least one of such features. In the description of this application, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are intended only to illustrate the relative positional relationships and movement of components in a specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly. Furthermore, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements and may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to such process, method, product, or apparatus.

[0025] References to "embodiments" herein mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0026] Example 1: Please refer to Figure 1 , a parameter optimization method for digital model generative design, the specific steps are as follows: S1. Initial model calibration: calibrate the basic digital model according to the design requirements and constraints of the project; S2. Global model establishment: conducting a global search, combining physical constraints and functional requirements, and conducting a multi-dimensional evaluation of the dam design scheme to generate a global model based on the basic digital model; S3. Local model establishment: using genetic algorithm combined with local search strategy, local model is established based on the global model, and the local model is analyzed and tested; S4, multi-objective optimization, uses multi-objective optimization algorithm to optimize and reanalyze the local model; S5. Feedback: Feedback each optimization parameter to the visualization terminal.

[0027] In this embodiment: Step S1: Initial model calibration, initial model calibration is the basic link of the entire parameter optimization process. Through a comprehensive review and integration of the design requirements and constraints of the project, this step performs targeted calibration of the basic digital model. This process ensures that the model is highly consistent with the actual engineering background, and improves the feasibility and scientificity of subsequent design solutions. The setting of design boundary conditions and the screening of key input variables are completed, laying a solid foundation for building high-quality global and local models. The main effect of this step is to significantly improve the accuracy and adaptability of the initial state of the model, and effectively reduce the risk of deviation in the subsequent optimization process.

[0028] Step S2: Global Model Construction. This phase integrates a global search algorithm to comprehensively scan and explore the space of possible design solutions. This process integrates the dam's physical constraints (such as structural stability and water pressure control) with functional requirements (such as water storage capacity and flood discharge performance) to conduct a multi-dimensional evaluation of various design options. The resulting global model has a strong coverage capability, reflecting the macroscopic characteristics of the entire solution space. This step shifts from single-solution thinking to systematic design exploration, providing a comprehensive reference and foundation for subsequent local refinement and optimization.

[0029] Step S3: Local Model Construction. Building on the global model, local model construction employs a genetic algorithm and a local search strategy to deeply explore potential high-quality solution intervals. This method combines the global nature of evolutionary mechanisms with the precision of local search, efficiently optimizing and analyzing local models and effectively avoiding local optima. Fine-tuning design parameters and sensitivity analysis further enhance the model's accuracy and practicality. Completion of this phase marks a transition from extensive exploration to precise modeling, significantly improving the efficiency and quality of solution optimization.

[0030] Step S4: Multi-Objective Optimization. This stage utilizes a multi-objective optimization algorithm to balance and coordinate optimization across multiple performance metrics. By reanalyzing local models and adjusting parameters, a set of design solutions with superior performance that meet practical multi-dimensional requirements is obtained. This step achieves an optimal balance between objectives such as cost, structural safety, and operational efficiency, while also enhancing the ability to solve complex engineering problems. It not only improves the comprehensiveness and robustness of the optimization results, but also enhances the operability and diversity of the design solutions in practical applications.

[0031] Step S5: Feedback and Visualization. In the final feedback phase, optimized parameters and design results are intuitively fed back to the visualization terminal, facilitating understanding and decision-making by designers. By displaying the optimization process and results through a graphical, interactive interface, engineers can quickly grasp key trends and rapidly adjust and iterate their solutions. This step significantly improves the transparency and interactivity of the design process, transforming it from "black-box" optimization to "interpretable optimization," enhancing the practicality and user experience of the entire system.

[0032] Compared to traditional engineering parameter optimization methods, such as manual adjustments based on experience or single-algorithm optimization, this parameter optimization process for generative digital model design introduces a multi-level modeling and multi-objective optimization mechanism, significantly improving the scientific and systematic nature of the design solution. Through a comprehensive process from initial model calibration to global and local modeling, and then to multi-objective trade-offs and result visualization, this method effectively integrates computing resources and design intelligence, achieving a "data-driven + experience-based" intelligent optimization path. This not only improves design efficiency, model accuracy, and solution quality, but also significantly reduces design risk and trial-and-error costs.

[0033] Example 2: Please refer to Figure 1 In step S1, the specific steps of calibrating the basic digital model are as follows: S1.1. Extract key design parameters based on the project data, specification requirements, and environmental conditions to generate structural load data sets, strength coordination data sets, and load-settlement parameter data sets. Standardize the generated data to ensure consistency between the data structure and the modeling interface. The structural load data group includes the dead load value YZ and the live load value RZ; The strength match data set includes the water-binder ratio value SJB and the cubic compressive strength value LFK; The bearing settlement parameter data group includes the foundation bearing capacity characteristic value DJT and the foundation deformation modulus value DJM; S1.2. Use digital modeling tools to construct a preliminary digital structural model. Input the processed structural load data set, strength coordination data set, and load-settlement parameter data set into the digital model to construct the initial calibration value JCZ. The specific calculation formula is as follows:

[0034] Where: YZ is the dead load value, RZ is the live load value, LFK is the cubic compressive strength value, SJB is the water-cement ratio value, DJT is the foundation bearing capacity characteristic value, and DJM is the foundation deformation modulus value.

[0035] In this embodiment, key design parameters such as structural load, strength coordination, and bearing settlement are standardized to ensure interface consistency and format standardization of each data group during the structural modeling process, thereby improving the systematicness and operability of the model input. The extraction of key design parameters is strictly based on the engineering data, relevant specifications, and environmental conditions provided by the project, effectively achieving an organic coupling between the actual constraints of the project and the modeling logic, making the constructed digital model more in line with the actual working conditions, and improving the engineering interpretation ability and practicality of the model results.

[0036] By introducing multidimensional parameters of structural load, strength performance, and foundation response, a unified initial calibration value JCZ calculation formula is constructed. This provides a clear quantitative benchmark for subsequent structural performance evaluation and optimization direction judgment, and has good reference and comparability. The initial calibration value JCZ calculation formula comprehensively considers the combined effects of dead load and live load, as well as the influence of foundation bearing capacity and deformation modulus on structural response. It can effectively reflect the coupling relationship between the structure-foundation system and improve the expression accuracy of the model under the influence of multiple factors such as complex loads and weak foundations. Through the systematic calibration and formulaic expression of initial parameters, the digital model is placed in a more reasonable engineering response range at the beginning of optimization, reducing invalid iterations caused by initial modeling deviations, effectively improving optimization efficiency, and reducing the waste of computing resources. The design variables involved in the initial calibration value JCZ indicator can serve as an important reference for objective function construction and parameter sensitivity evaluation in the subsequent optimization process. It has clear physical meaning and engineering relevance, providing data support for the convergence and robustness of the multi-objective optimization algorithm.

[0037] Example 3: Please refer to Figure 1 ,In step S2, the specific steps of establishing the global model are as follows: Step S2.1: Collect the design goals for the safety, economy, and functional integrity of the dam project, construct a functional requirement data set, and construct a physical constraint data set based on the project specifications and operating conditions. Standardize the generated data to ensure that the data structure is consistent with the modeling interface. The functional requirement data group includes the effective storage capacity value KRZ, the downstream flow regulation capacity value XHZ and the structural stiffness value GTZ; The physical constraint data set includes the anti-sliding stability safety factor AQZ, the maximum allowable displacement value of the dam body WYZ, and the maximum compressive stress value of the dam foundation YLZ; Step S2.2: Based on the established multi-objective optimization model, genetic algorithm and particle swarm optimization are used to iterate the initial calibration value JCZ according to the functional requirement data set and the physical constraint data set in the multi-dimensional parameter space to generate the global calibration value QJY. The specific calculation formula is as follows:

[0038] Where: KRZ is the effective storage capacity, XHZ is the discharge flow regulation capacity, GTZ is the structural stiffness, AQZ is the anti-sliding stability safety factor, WYZ is the maximum allowable displacement of the dam body, YLZ is the maximum compressive stress of the dam foundation, and JCZ is the initial calibration value.

[0039] In this embodiment: by constructing functional requirement data groups and physical constraint data groups separately and standardizing them, the data barriers between structural functional performance indicators and physical constraint boundaries are broken down, and for the first time, the coordinated expression of engineering performance goals and safety control parameters at the model level is achieved, providing a comprehensive decision-making basis for optimized design.

[0040] Relying on the joint iterative mechanism of genetic algorithm and particle swarm optimization, the problems of high dimension of parameter space and strong nonlinearity of objective function are effectively solved, enabling the model to quickly converge to the global optimal solution range under complex design objectives and constraints, significantly improving the depth and breadth of the model's search for the solution space, and avoiding falling into local optimality.

[0041] All variables in the global calibration value QJY formula are derived from key response indicators under different working conditions. Therefore, the model can maintain stable response capabilities and optimization directions when dealing with different operating states, effectively improving the engineering adaptability and safety redundancy capabilities of the design scheme.

[0042] By setting clear mathematical relationships and parameter transfer paths, the evolutionary logic of the optimization process is made explainable and traceable, providing engineering technicians with a clear data basis and path control capabilities for subsequent solution adjustments, sensitivity analysis, and parameter backtracking.

[0043] The model performs well in balancing the multiple objectives of "safety, economy, and functional integrity." The global calibration value QJY calculated and output by the global model can be effectively used as a comprehensive evaluation indicator for optimization and screening, significantly improving the design team's selection efficiency and engineering feasibility among multiple candidate solutions.

[0044] Example 4: Please refer to Figure 1 In step S3, the specific steps of establishing the local model are as follows: S3.1. Collect stability and failure boundary parameters, construct optimization solution verification data sets, stability data sets, and failure boundary data sets, and standardize the generated data to ensure that the data structure is consistent with the modeling interface; The optimization solution verification data set includes deformation BXL and deformation stiffness BXG; The stability data set includes anti-slip force KHL and sliding force HDL; The failure boundary data set includes fracture toughness DLR and ultimate load JHZ; S3.2. By coupling the stability data set and the failure boundary data set with the optimization solution verification data set and the global calibration value QJY, respectively, a genetic algorithm is used to perform local iterative updates on the stability and failure boundaries to synchronously adjust multiple data sets, thereby generating a stability optimization reference coefficient WDX and a failure boundary optimization reference coefficient PBJ; S3.3. Analyze the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ to determine the local optimization range.

[0045] In this embodiment: by collecting and standardizing the key parameters of stability and failure boundaries, a multi-dimensional input system including an optimization solution verification data group, a stability data group, and a failure boundary data group is constructed, thereby effectively achieving local refined processing of key areas of the global model and providing more targeted analysis methods for complex stress concentration areas.

[0046] This step logically couples the stability and damage boundary data groups with the optimization solution verification data group and the global calibration value QJY, respectively, and thus realizes the response inheritance and adaptation in the process of transforming the overall structural performance indicators into the local safety boundary for the first time, thereby ensuring that the optimization adjustment satisfies both global performance and local structural safety.

[0047] The coupling parameters are iteratively optimized through genetic algorithms, and the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ are constructed respectively. This realizes the synchronous adjustment of multiple sets of local key parameters under nonlinear response relationships, effectively improving the reliability and stability of the optimization solution.

[0048] As composite response indicators of stability and failure boundaries, WDX and PBJ can effectively reflect problems such as slippage, cracking and bearing limit that may occur in local areas of the dam body. They provide a quantitative basis for early identification of high-risk components or structural weaknesses, and enhance structural health monitoring and risk management capabilities.

[0049] By analyzing the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ, the model can automatically identify local structural areas that require further refinement or adjustment, accurately define the optimization range, avoid resource waste or redundant calculations of non-critical parts during the global optimization process, and thus improve optimization efficiency and calculation accuracy.

[0050] The stability optimization reference coefficient WDX and failure boundary optimization reference coefficient PBJ output from the local optimization results can not only be used to evaluate structural responses, but also serve as quantitative indicators for formulating structural reinforcement strategies, adjusting design redundancy, or evaluating material performance requirements, providing practical decision-making support for subsequent project implementation.

[0051] Example 5: Please refer to Figure 1In step S3.2, the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ are calculated as follows:

[0052]

[0053] Where: BXL is the deformation, BXG is the deformation stiffness, KHL is the anti-sliding force, HDL is the sliding force, DLR is the fracture toughness, JHZ is the ultimate load, and QJY is the global calibration value; a1 and a2 are empirical adjustment factors. The value range of a1 is [0.5, 2], and the value range of a2 is [0.5, 3]. The specific values ​​of a1 and a2 are adjusted by the user. b1 and b2 are nonlinear exponents. The value range of b1 is [0.3, 1.5], and the value range of b2 is [0.5, 2.5]. The specific values ​​of b1 and b2 are adjusted and set by the user.

[0054] In this example, by exponentially coupling deformation and deformation stiffness with key parameters such as the structural anti-sliding force, sliding force, fracture toughness, and ultimate load, combined with empirical adjustment factors, a refined, nonlinear modeling of local structural performance is achieved. This improvement makes the structure's response to complex loads and ultimate limit states more consistent with actual working conditions, avoiding errors caused by linear simplification.

[0055] By introducing user-adjustable empirical adjustment factors, the model can be flexibly adjusted to meet performance requirements under different working conditions based on different engineering needs or design preferences. This flexibility allows the optimization process to adapt to different design standards or site conditions, enhancing the system's adaptability and operability in changing environments.

[0056] The specific value ranges for a1, a2, b1, and b2 provide users with reasonable adjustment space, enabling precise local optimization adjustments to different design solutions while ensuring structural safety and stability. This design not only ensures high theoretical accuracy of the model, but also meets engineers' intuitive understanding and design needs in actual operations.

[0057] By normalizing the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ with the global calibration value QJY, the coordination between local optimization results and global performance is further ensured. This structured, multi-level calibration mechanism effectively avoids conflicts between global optimization and local refinement, providing more reliable and consistent performance predictions throughout the design process.

[0058] The stability optimization reference coefficient (WDX) and the failure boundary optimization reference coefficient (PBJ) serve as optimization coefficients for stability and failure boundaries, dynamically reflecting the stability and critical failure states of the structure under different loads and operating conditions. Using these coefficients, the design team can assess the dam's safety redundancy and risk points in real time during the optimization process, proactively identifying potential local instability or structural failure risks and providing a scientific basis for further design adjustments or reinforcement.

[0059] Example 6: Please refer to Figure 1 In step S3.3, the specific analysis of the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ is as follows: when When , it means that there is no abnormality in the digital model in terms of structural positioning and no optimization is required; when When , it means that the digital model has an abnormality in structural positioning and needs to be optimized; when When , it means that there is no abnormality in the digital model on the damage boundary and no optimization is required; when , it means that the digital model is abnormal on the failure boundary and needs to be optimized.

[0060] In this embodiment, by setting thresholds for the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ, a set of quantifiable and clear structural performance judgment criteria is established. This mechanism effectively identifies potential anomalies in the model's structural positioning and damage boundary, achieving a shift from "empirical judgment" to "data-driven judgment," enhancing the scientific nature and objectivity of the structural optimization process.

[0061] Through real-time calculation and comparison of WDX and PBJ, the model can quickly determine whether the current structure has performance defects that need to be adjusted in a local range, thereby quickly locking in the key optimization areas in the multi-dimensional parameter space, avoiding invalid calculations and waste of resources, and significantly improving optimization efficiency and judgment accuracy.

[0062] This step embeds judgment logic into the optimization process. When an anomaly is detected, the corresponding local optimization process is automatically triggered, forming a closed-loop feedback control mechanism of "calculation-judgment-feedback-adjustment." This mechanism significantly enhances the model's adaptive optimization capabilities when structural responses are abnormal, enabling intelligent triggering and dynamic response of structural optimization decisions.

[0063] The above-mentioned criteria enable the digital model to have the functions of self-verification and self-diagnosis. It can not only identify anomalies, but also clarify the types of anomalies, providing optimization personnel with highly targeted intervention directions and reducing the risks of human misjudgment and repeated modeling. Example 7: Please refer to Figure 1 In step S4, the specific steps of multi-objective model optimization are as follows: S4.1. Collect supplementary parameters, construct a supplementary data set for structural stability and an extended data set for failure behavior, and standardize the generated data to ensure that the data structure is consistent with the modeling interface; The stability supplementary data set includes support spacing ZCJ, lateral displacement CYW and horizontal equivalent stiffness SPG; The extended data set of failure behavior includes crack length LFC, peak stress FYL, and crack tip radius LJB; S4.2. Use a genetic algorithm to perform local iterative updates on the structural stability supplementary data set and the failure behavior extended parameter set, respectively, with the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ to synchronously adjust multiple data. The structural stability supplementary data set and the failure behavior extended parameter set are coupled with the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ to generate the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB. S4.3. Analyze the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB to locate the defect area of ​​the model.

[0064] In this example, by collecting supplementary parameters and constructing a supplementary data set for structural stability and an extended data set for failure behavior, the model not only considers traditional static stability and failure boundaries, but also incorporates more complex factors such as dynamic responsiveness and crack evolution. This expansion enables the model to more comprehensively and accurately reflect the long-term operational behavior and potential damage of the dam under various operating conditions, significantly enhancing the model's adaptability and predictive accuracy.

[0065] A genetic algorithm is used to iteratively update the supplementary stability data set, the extended failure behavior data set, and the optimized reference coefficients WDX and PBJ locally, ensuring synchronized adjustments within the multidimensional parameter space. This generates the dynamic structural location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB. This method automatically seeks the optimal solution during the optimization process, making each parameter update more precise and avoiding the problem of inadequate local optimization affecting the overall model accuracy.

[0066] By generating dynamic structural location coefficients (DWD) and dynamic failure boundary optimization location coefficients (DPB), the model can dynamically adjust design parameters based on real-time calculation results, flexibly optimizing design solutions in response to varying environmental conditions. This dynamic adaptability greatly enhances the model's robustness and real-time responsiveness under complex working conditions.

[0067] In traditional structural optimization, static parameter adjustment is common. However, this step realizes dynamic iteration and optimization during the design process by introducing the dynamic structural positioning coefficient DWD and the failure boundary optimization positioning coefficient DPB. This makes the structural performance analysis not only limited to a fixed state, but can adjust parameters in real time as different working conditions change, ensuring that the optimization results meet the needs of different stages and environments.

[0068] By analyzing the dynamic structural location coefficient (DWD) and the dynamic failure boundary optimization location coefficient (DPB), potential defect areas in the model can be effectively identified. Especially in complex designs, dynamic optimization can accurately locate possible structural weaknesses or damage development trends, providing a basis for subsequent reinforcement solutions and structural optimization.

[0069] The multi-objective optimization model in step S4 takes into account multiple objectives, including structural stability, failure boundaries, and dynamic behavior. By comprehensively optimizing each objective and performing local iterative adjustments, the resulting optimization solution maximizes the balance of various design requirements, enhancing the scientific nature and practical operability of the design and avoiding the design deviations and imperfections that may arise from single-objective optimization.

[0070] Example 8: Please refer to Figure 1 In step S4.2, the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB are calculated as follows:

[0071]

[0072] Where: ZCJ is the support spacing, CYW is the lateral displacement, SPG is the horizontal equivalent stiffness, LFC is the crack length, FYL is the peak stress, LJB is the crack tip radius, WDX is the stability optimization reference coefficient, and PBJ is the failure boundary optimization reference coefficient.

[0073] In this embodiment, by introducing typical structural response parameters, the calculated dynamic structural location coefficient DWD and dynamic failure boundary optimization location coefficient DPB can more precisely reflect the mechanical behavior and damage evolution process of the structure in the local area, so that the model has the ability to sensitively identify the evolution of complex structural states.

[0074] The calculation method adopted exponentially couples the structural response parameters with the pre-optimization reference coefficients to construct a nonlinear enhancement function.

[0075] The dynamic structural location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB start from the two directions of structural stability and failure behavior respectively, and form a composite response coefficient through the interaction between multiple parameters. They can accurately identify the weakening of structural performance caused by local stiffness attenuation, displacement anomalies or crack extension, and significantly improve the consistency and matching between model response prediction and structural diagnosis.

[0076] Traditional optimization often judges optimization needs based on a single indicator, while the dynamic structure positioning coefficient DWD and dynamic damage boundary optimization positioning coefficient DPB proposed in this step drive the optimization adjustment process through dynamic variables, realizing the transition from static threshold judgment to dynamic response mapping, effectively improving the model's adaptability in a dynamic environment. The dynamic structure positioning coefficient DWD and dynamic damage boundary optimization positioning coefficient DPB are not only used to locate structural abnormal areas, but can also be used as highly sensitive performance indicators in a digital simulation environment to track and quantify potential risk points in the structural response process, ultimately achieving a closed-loop integration of design-simulation-optimization, and promoting digital modeling to develop in the direction of integrated structural safety assessment. Through quantitative evaluation of the dynamic structure positioning coefficient DWD and the dynamic damage boundary optimization positioning coefficient DPB, not only can the rationality of the current optimization state be identified, but a reliable data basis can also be provided for subsequent multi-objective constraint optimization, thereby improving the scientific nature and execution of the overall optimization path.

[0077] Example 9: Please refer to Figure 1 In step S4.3, the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB are specifically analyzed as follows: when When , it means that the accuracy of digital model generation in structural positioning is at the first level and no adjustment is required; when When , it means that the accuracy of digital model generation in structural positioning is at the second level, and the structural positioning model needs to be locally optimized; when When , it means that the accuracy of digital model generation in structural positioning is at level 3, and the structural positioning model needs to be fully optimized; when When , it means that the accuracy of digital model generation on the damage boundary is at the first level and no adjustment is required; when , which means that the accuracy of digital model generation on the damage boundary is in the secondary state, and the damage boundary model needs to be locally optimized; when When , it means that the accuracy of digital model generation on the damage boundary is at level 3, and the damage boundary model needs to be fully optimized.

[0078] In this embodiment, by classifying the dynamic structural location coefficient (DWD) and the dynamic failure boundary optimization location coefficient (DPB) into three levels of evaluation criteria, the accuracy of the digital model in terms of structural location and failure boundary can be quantitatively judged. This hierarchical evaluation mechanism makes the model optimization process more scientific and controllable, eliminating reliance on empirical judgment, thereby enhancing the targetedness and effectiveness of the overall optimization.

[0079] Automatically triggering local or global optimization strategies based on the dynamic structural location coefficient (DWD) and the dynamic failure boundary optimization location coefficient (DPB) within different intervals enables real-time response and dynamic adjustment to changes in model performance. In particular, when model performance degrades, the system promptly identifies issues and guides optimization paths, helping to reduce error accumulation and improve model stability and reliability.

[0080] For the model part with first-level accuracy, no optimization is required, which can save computing resources and development costs; for the part with second-level or third-level accuracy, local or comprehensive optimization strategies are adopted respectively to achieve reasonable resource allocation and maximize efficiency.

[0081] Separately evaluating the performance of structural positioning and damage boundaries helps identify specific bottlenecks in model optimization. For example, when the structure is accurate but the boundary deviation is large, only the boundary module needs to be optimized, avoiding unnecessary overall adjustments. This improves optimization efficiency and module independence. By using clear optimization metrics and standards, it helps to continuously improve the performance of digital models in complex structure identification and damage boundary delineation, thereby providing higher-quality data support for subsequent engineering simulation, risk assessment, and decision-making.

[0082] In addition, the functional units in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the description and drawings of this application, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

[0083] The above detailed description of the specific embodiments of the invention is intended only as an example, and the present application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions of the invention are also within the scope of the present application. Therefore, equivalent changes, modifications, and improvements made without departing from the spirit and scope of the present application should be included within the scope of the present application.

Claims

1. A parameter optimization method for generative design of digital models, characterized by: The specific steps are as follows: S1. Initial model calibration: calibrate the basic digital model according to the design requirements and constraints of the project; S2. Global model establishment: conducting a global search, combining physical constraints and functional requirements, and conducting a multi-dimensional evaluation of the dam design scheme to generate a global model based on the basic digital model; S3. Local model establishment: using genetic algorithm combined with local search strategy, local model is established based on the global model, and the local model is analyzed and tested; S4, multi-objective optimization, uses multi-objective optimization algorithm to optimize and reanalyze the local model; S5. Feedback: Feedback each optimization parameter to the visualization terminal.

2. The parameter optimization method for generative design of digital models according to claim 1, characterized in that: In step S1, the specific steps of calibrating the basic digital model are as follows: S1.

1. Extract key design parameters based on the project data, specification requirements, and environmental conditions to generate structural load data sets, strength coordination data sets, and load-settlement parameter data sets. Standardize the generated data to ensure consistency between the data structure and the modeling interface. The structural load data group includes the dead load value YZ and the live load value RZ; The strength match data set includes the water-binder ratio value SJB and the cubic compressive strength value LFK; The bearing settlement parameter data group includes the foundation bearing capacity characteristic value DJT and the foundation deformation modulus value DJM; S1.

2. Use digital modeling tools to construct a preliminary digital structural model. Input the processed structural load data set, strength coordination data set, and load-settlement parameter data set into the digital model to construct the initial calibration value JCZ. The specific calculation formula is as follows: Where: YZ is the dead load value, RZ is the live load value, LFK is the cubic compressive strength value, SJB is the water-cement ratio value, DJT is the foundation bearing capacity characteristic value, and DJM is the foundation deformation modulus value.

3. The parameter optimization method for generative design of digital models according to claim 2, characterized in that: In step S2, the specific steps of establishing the global model are as follows: Step S2.1: Collect the design goals for the safety, economy, and functional integrity of the dam project, construct a functional requirement data set, and construct a physical constraint data set based on the project specifications and operating conditions. Standardize the generated data to ensure that the data structure is consistent with the modeling interface. The functional requirement data group includes the effective storage capacity value KRZ, the downstream flow regulation capacity value XHZ and the structural stiffness value GTZ; The physical constraint data set includes the anti-sliding stability safety factor AQZ, the maximum allowable displacement value of the dam body WYZ, and the maximum compressive stress value of the dam foundation YLZ; Step S2.2: Based on the established multi-objective optimization model, genetic algorithm and particle swarm optimization are used to iterate the initial calibration value JCZ according to the functional requirement data set and the physical constraint data set in the multi-dimensional parameter space to generate the global calibration value QJY. The specific calculation formula is as follows: Where: KRZ is the effective storage capacity, XHZ is the discharge flow regulation capacity, GTZ is the structural stiffness, AQZ is the anti-sliding stability safety factor, WYZ is the maximum allowable displacement of the dam body, YLZ is the maximum compressive stress of the dam foundation, and JCZ is the initial calibration value.

4. The parameter optimization method for generative design of digital models according to claim 3, characterized in that: In step S3, the specific steps of establishing the local model are as follows: S3.

1. Collect stability and failure boundary parameters, construct optimization solution verification data sets, stability data sets, and failure boundary data sets, and standardize the generated data to ensure that the data structure is consistent with the modeling interface; The optimization solution verification data set includes deformation BXL and deformation stiffness BXG; The stability data set includes anti-slip force KHL and sliding force HDL; The failure boundary data set includes fracture toughness DLR and ultimate load JHZ; S3.

2. By coupling the stability data set and the failure boundary data set with the optimization solution verification data set and the global calibration value QJY, respectively, a genetic algorithm is used to perform local iterative updates on the stability and failure boundaries to synchronously adjust multiple data sets, thereby generating a stability optimization reference coefficient WDX and a failure boundary optimization reference coefficient PBJ; S3.

3. Analyze the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ to determine the local optimization range.

5. The parameter optimization method for generative design of digital models according to claim 4, characterized in that: In step S3.2, the stability optimization reference coefficient WDX and the damage boundary optimization reference coefficient PBJ are calculated as follows: Where: BXL is the deformation, BXG is the deformation stiffness, KHL is the anti-sliding force, HDL is the sliding force, DLR is the fracture toughness, JHZ is the ultimate load, and QJY is the global calibration value; a1 and a2 are empirical adjustment factors. The value range of a1 is [0.5, 2], and the value range of a2 is [0.5, 3]. The specific values ​​of a1 and a2 are adjusted by the user. b1 and b2 are nonlinear exponents. The value range of b1 is [0.3, 1.5], and the value range of b2 is [0.5, 2.5]. The specific values ​​of b1 and b2 are adjusted and set by the user.

6. The parameter optimization method for generative design of digital models according to claim 5, characterized in that: In step S3.3, the specific analysis of the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ is as follows: when When , it means that there is no abnormality in the digital model in terms of structural positioning and no optimization is required; when When , it means that the digital model has an abnormality in structural positioning and needs to be optimized; when When , it means that there is no abnormality in the digital model on the damage boundary and no optimization is required; when , it means that the digital model is abnormal on the damage boundary and needs to be optimized.

7. The parameter optimization method for generative design of digital models according to claim 6, characterized in that: In step S4, the specific steps of multi-objective model optimization are as follows: S4.

1. Collect supplementary parameters, construct a supplementary data set for structural stability and an extended data set for failure behavior, and standardize the generated data to ensure that the data structure is consistent with the modeling interface; The stability supplementary data set includes support spacing ZCJ, lateral displacement CYW and horizontal equivalent stiffness SPG; The extended data set of failure behavior includes crack length LFC, peak stress FYL, and crack tip radius LJB; S4.

2. Use a genetic algorithm to perform local iterative updates on the structural stability supplementary data set and the failure behavior extended parameter set, respectively, with the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ to synchronously adjust multiple data. The structural stability supplementary data set and the failure behavior extended parameter set are coupled with the stability optimization reference coefficient WDX and the failure boundary optimization reference coefficient PBJ to generate a dynamic structural location coefficient DWD and a dynamic failure boundary optimization location coefficient DPB. S4.

3. Analyze the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB to locate the defect area of ​​the model.

8. The parameter optimization method for generative design of digital models according to claim 7, characterized in that: In step S4.2, the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB are calculated as follows: Where: ZCJ is the support spacing, CYW is the lateral displacement, SPG is the horizontal equivalent stiffness, LFC is the crack length, FYL is the peak stress, LJB is the crack tip radius, WDX is the stability optimization reference coefficient, and PBJ is the failure boundary optimization reference coefficient.

9. The parameter optimization method for generative design of digital models according to claim 8, characterized in that: In step S4.3, the dynamic structure location coefficient DWD and the dynamic failure boundary optimization location coefficient DPB are specifically analyzed as follows: when When , it means that the accuracy of digital model generation in structural positioning is at the first level and no adjustment is required; when When , it means that the accuracy of digital model generation in structural positioning is at the second level, and the structural positioning model needs to be locally optimized; when When , it means that the accuracy of digital model generation in structural positioning is at level 3, and the structural positioning model needs to be fully optimized; when When , it means that the accuracy of digital model generation on the damage boundary is at the first level and no adjustment is required; when , which means that the accuracy of digital model generation on the damage boundary is in the secondary state, and the damage boundary model needs to be locally optimized; when When , it means that the accuracy of digital model generation on the damage boundary is at level 3, and the damage boundary model needs to be fully optimized.

Images