A method and device for generating an initial configuration of a monopile foundation for offshore wind power and a medium

Abstract

The present application relates to the offshore wind power technical field, disclose a kind of offshore wind power single pile foundation initial configuration generation method, device and medium, the present application obtains the engineering parameter and soil layer parameter of target position;Based on engineering parameter and soil layer parameter, construct the constraint set for constraining single pile foundation geometric parameter;Based on constraint set, using first generation path or second generation path generates candidate parameter set, wherein the first generation path can be quickly sampled in the constraint feasible region to obtain the feasible solution satisfying basic constraint, the second generation path can utilize large model to generate more close to engineering actual candidate parameter by automatically modeling complex soil layer characteristics and multidimensional constraint relationship, based on candidate parameter set, the geometric parameter of each pile segment of single pile foundation is generated by geometric principle, as the initial configuration of single pile foundation.The present application effectively gets rid of the dependence on artificial experience repeated iteration, significantly improves the design efficiency and rationality of single pile foundation initial configuration.

Description

Technical Field

[0001] This invention relates to the field of offshore wind power technology, specifically to a method, device, and medium for generating the initial configuration of an offshore wind turbine monopile foundation. Background Technology

[0002] Offshore wind power is a key area of ​​global clean energy development, and its core load-bearing structure is the foundation system of the offshore wind turbine. Among these, monopile foundations have become the most widely used foundation type due to their relatively mature construction technology, controllable manufacturing costs, short construction period, and wide applicable water depth range. During service, monopile foundations need to withstand complex marine environmental loads such as wind, waves, and currents, as well as the cyclic forces generated by the operation of the upper wind turbine. Their long-term stability and safety directly depend on the structural parameters selected in the initial design stage.

[0003] In practical engineering, the structural parameters of monopile foundations (such as pile diameter, wall thickness, pile length, and penetration depth) not only affect the overall structure's bearing capacity and fatigue resistance, but are also closely related to construction feasibility, material usage, and the economics throughout the entire life cycle. Inappropriate parameter design can lead to overly conservative foundation design, increasing costs, or insufficient bearing capacity, resulting in structural risks. Therefore, how to reasonably and quickly obtain an initial configuration that meets specifications and engineering requirements under complex constraints has become a key issue in offshore wind power design.

[0004] In existing technologies, the initial configuration design of monopile foundations mainly relies on the accumulated experience of designers and the calculation methods of current standards. Designers often need to determine the combination of key parameters such as pile diameter, wall thickness, and embedment depth through repeated iterative calculations and manual judgment under multiple constraints, including complex marine geological conditions, wind, wave and current coupling effects, construction equipment capabilities, and safety regulations. Due to the numerous and coupled types of constraints, the traditional manual design process usually requires extensive calculations and comparisons under multiple assumed working conditions, which is time-consuming and the results are highly dependent on the designer's subjective experience.

[0005] Furthermore, existing methods struggle to fully account for multidisciplinary coupling effects in the initial stages, such as foundation nonlinear response, fatigue life assessment, and construction technology limitations, often resulting in limitations in the initial configuration. If the initial parameters are not reasonable enough, subsequent engineering simulations and optimization calculations require significant adjustments and corrections, thus significantly increasing the design cycle and computational costs. In the context of large-scale offshore wind farm development, this inefficient manual initial design method is no longer sufficient to meet the demands of rapid, batch, and high-precision engineering. Summary of the Invention

[0006] This invention provides a method, device, and medium for generating initial configurations of offshore wind power monopile foundations that can automatically generate initial configurations that satisfy constraint sets and take into account both sampling and large model dual-path generation. This solves the problems of existing monopile foundation initial configuration design relying on manual experience, low efficiency, and difficulty in taking into account multiple constraint conditions.

[0007] In a first aspect, the present invention provides a method for generating the initial configuration of a monopile foundation for offshore wind power, comprising: acquiring engineering parameters and soil parameters of the target turbine location; constructing a constraint set for constraining the geometric parameters of the monopile foundation based on the engineering parameters and soil parameters; generating a candidate parameter set based on the constraint set using a first generation path or a second generation path; wherein the first generation path is: sampling within the feasible domain defined by the constraint set; the second generation path is: generating candidate parameters through a large model based on the engineering parameters, soil parameters, and constraint set, and normalizing the candidate parameters generated by the large model so that they fall within the range defined by the constraint set; and generating the geometric parameters of each pile segment of the monopile foundation based on the candidate parameter set using geometric principles, as the initial configuration of the monopile foundation.

[0008] This invention constructs a constraint set and employs a dual-path generation mechanism with a first generation path and a second generation path running in parallel. The first generation path can quickly sample within the feasible domain of constraints to obtain feasible solutions that satisfy the basic constraints, while the second generation path can automatically model complex soil characteristics and multi-dimensional constraint relationships using a large model to generate candidate parameters that are closer to engineering reality. The two paths can be used selectively or mutually as a fallback based on preset conditions, thereby effectively eliminating the reliance on repeated iterations of manual experience, significantly improving the design efficiency and rationality of the initial configuration of the single pile foundation, and providing better initial input for subsequent simulation optimization.

[0009] In one optional implementation, the engineering parameters include at least one of the following: flange top elevation, mud surface elevation, mean sea level elevation, target mud penetration depth, maximum pile diameter, and diameter-to-thickness ratio range; the soil layer parameters include at least one of the following: soil layer top elevation, soil layer bottom elevation, soil type, buoyant unit weight, internal friction angle, and undrained shear strength; the constraint set includes at least one of the following: maximum diameter constraint, diameter-to-thickness ratio constraint, tapered transition section slope constraint, and minimum section length constraint. This implementation, by clearly defining the specific composition of the engineering parameters, soil layer parameters, and constraint set, ensures the integrity of the input data and the quantifiability of the constraints, thereby providing a unified and standardized data foundation and clear optimization boundaries for subsequent dual-path generation and geometric solution. This effectively avoids design iterations and unreasonable configurations caused by missing parameters or ambiguous constraints, further improving the generation efficiency and reliability of the initial configuration of the single pile foundation.

[0010] In one optional implementation, a candidate parameter set is generated using either a first generation path or a second generation path, including: using the second generation path when preset conditions are met; and using the first generation path when preset conditions are not met or the second generation path is abnormal. This implementation achieves intelligent switching between the large model generation path and the sampling path by establishing a path selection mechanism based on preset conditions and anomaly detection: when conditions are met, the large model is preferentially used to generate candidate parameters that are closer to the actual engineering situation; and when conditions are insufficient or the large model is abnormal, it automatically reverts to the sampling path to ensure the rapid acquisition of basic feasible solutions. This fully leverages the reasoning capabilities of the large model while ensuring the robustness and availability of the solution under various working conditions, avoiding the failure or unusable output of a single path in specific scenarios.

[0011] In one optional implementation, sampling is performed within the feasible region defined by the constraint set, including: sampling using at least one of Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling within the feasible region defined by the constraint set. This implementation achieves efficient coverage and uniform sampling of the high-dimensional parameter space by employing at least one of Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling within the feasible region defined by the constraint set. It can more comprehensively explore the boundary of the feasible region with a limited number of samplings, avoiding the blindness and local clustering problems of traditional random sampling. This provides a more representative and diverse set of candidate parameters for subsequent geometric solutions, improving the global search capability and optimal solution space for initial configuration generation.

[0012] In one optional implementation, candidate parameters are generated through a large model based on engineering parameters, soil layer parameters, and constraint sets. This includes: constructing prompts containing output format requirements based on the engineering parameters, soil layer parameters, and constraint sets; inputting the prompts into the large model; and obtaining the candidate parameters generated by the large model according to the output format requirements. This implementation achieves a structured connection between the large model output and engineering specifications by constructing prompts containing output format requirements based on engineering parameters, soil layer parameters, and constraint sets, and inputting these prompts into the large model to obtain candidate parameters generated according to the output format requirements. By constraining the originally unstructured natural language reasoning process to a predefined output format, it effectively solves the common problems of uncontrollable format, missing fields, and parsing difficulties in engineering parameter generation scenarios, ensuring that the model output results can be directly and stably read and processed by the solver, significantly improving the overall automation integration capability and engineering feasibility of the technical solution.

[0013] In one optional implementation, the candidate parameters generated by the large model are normalized to fall within the range defined by the constraint set. This includes: converting the candidate parameters generated by the large model into standard units of measurement; and pruning parameters that exceed the boundary of the constraint set to their boundary values. This implementation, by performing unit normalization and boundary pruning on the candidate parameters generated by the large model, forcibly constrains the output of the large model, which might otherwise deviate from engineering reality, to a preset feasible range. This effectively solves the unit ambiguity and numerical out-of-bounds problems commonly encountered in engineering applications of large models, ensuring the physical reliability and computational feasibility of the candidate parameters in subsequent geometric solutions. It provides a crucial safety mechanism for the engineering implementation of large model generation paths.

[0014] In one optional implementation, the monopile foundation includes an upper straight section, a conical transition section, and a lower straight section, with the conical transition section located above the mud surface. Geometric parameters for each pile segment of the monopile foundation are generated using geometric principles, including: calculating the length, top elevation, and bottom elevation of the conical transition section based on parameters from a candidate parameter set; calculating the length of the upper straight section based on the flange top elevation and the top elevation of the conical transition section; and calculating the length of the lower straight section based on the bottom elevation of the conical transition section, the mud surface elevation, and the target mud penetration depth. This implementation establishes a three-segment geometric configuration comprising an upper straight section, a conical transition section, and a lower straight section. Based on the conical transition section parameters from a candidate parameter set, as well as fundamental data such as the flange top elevation, mud surface elevation, and target mud penetration depth, the length of the conical transition section, the elevation of each segment, and the lengths of the upper and lower straight sections are solved sequentially according to geometric principles, achieving automated and accurate conversion from discrete candidate parameters to continuous, manufacturable pile segment geometry. This deterministic solution process does not rely on manual trial calculations or empirical judgments, and can ensure strict continuity and physical consistency between the geometric parameters of each segment, providing a standardized and directly usable initial geometric model for subsequent structural simulation and detailed design.

[0015] In one optional implementation, after generating the geometric parameters of each pile segment of the monopile foundation, the method further includes: calculating the cumulative subsurface length of each pile segment; calculating the deviation between the cumulative subsurface length and the target subsurface depth; and when the absolute value of the deviation exceeds a preset tolerance, adjusting the length of the lower straight segment and regenerating the geometric parameters. This implementation introduces a closed-loop verification and callback mechanism for subsurface depth after the geometric parameters are generated. It automatically calculates the deviation between the cumulative subsurface length and the target subsurface depth, and actively adjusts the length of the lower straight segment and regenerates the geometric parameters when the deviation exceeds the preset tolerance requirement, until the preset tolerance requirement is met. This mechanism effectively solves the problem of cumulative subsurface depth error that may occur during parameter transfer and geometric derivation in multi-segment configurations. It ensures that the final output monopile foundation configuration accurately meets the engineering preset value for the key design indicator of subsurface depth, avoiding subsequent simulation errors or manual correction work caused by depth deviation, and further improving the engineering usability and automation of the generated results.

[0016] Secondly, the present invention provides an initial configuration generation device for offshore wind power monopile foundations, comprising: a data acquisition module for acquiring engineering parameters and soil parameters of the target turbine location; a constraint construction module for constructing a constraint set for constraining the geometric parameters of the monopile foundation based on the engineering parameters and soil parameters; a dual-path generation module for generating a candidate parameter set based on the constraint set using a first generation path or a second generation path; wherein the first generation path is: sampling within the feasible region defined by the constraint set; the second generation path is: generating candidate parameters through a large model based on the engineering parameters, soil parameters, and constraint set, and normalizing the candidate parameters generated by the large model so that they fall within the range defined by the constraint set; and a geometric solution module for generating the geometric parameters of each pile segment of the monopile foundation based on the candidate parameter set using geometric principles, as the initial configuration of the monopile foundation.

[0017] Thirdly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for generating the initial configuration of an offshore wind power monopile foundation according to the first aspect or any corresponding embodiment described above. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the first process for generating the initial configuration of a monopile foundation for offshore wind power according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the second process for generating the initial configuration of an offshore wind turbine monopile foundation according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating the collaborative process between the large model generation path and the deterministic verifier in an embodiment of the present invention. Figure 4 This is a schematic diagram of the three-section monopile configuration and its main geometric parameters in an embodiment of the present invention; Figure 5 This is a flowchart of the strong verification and error vector feedback in an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the interface between the unified table and the simulation in this embodiment of the invention; Figure 7This is a structural block diagram of an offshore wind power monopile foundation initial configuration generation device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0022] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0023] In existing technologies, the initial configuration design of monopile foundations mainly relies on the experience of designers and the calculation methods of current specifications. It requires repeated iterative calculations under multiple constraints such as marine geological conditions, wind, wave and current loads, construction equipment capabilities, and safety regulations. This process is time-consuming and the results are highly dependent on subjective experience. At the same time, it is difficult to fully consider the coupling effects of multiple disciplines in the initial stage, resulting in insufficient rationality of the initial configuration and a large workload for subsequent simulation optimization and adjustment. This makes it difficult to meet the needs of large-scale wind farm development for rapid, batch, and high-precision design.

[0024] With the development of artificial intelligence and large-scale modeling technology, using large-scale models to generate parameters for offshore wind turbine monopile foundations has become a new technical approach. Large-scale models possess the ability to integrate interdisciplinary knowledge and model complex relationships, and can quickly output initial configurations that meet requirements after inputting basic marine environmental parameters (such as water depth, soil layer distribution, and wind, wave, and current load characteristics) and engineering constraints. This method can significantly simplify the manual design process, improve the automation and intelligence level of the design phase, and provide better initial solutions for subsequent engineering simulation and optimization calculations, thereby accelerating design iteration and improving the rationality and engineering feasibility of structural schemes. Based on this, this invention aims to provide a method, device, and medium for generating initial configurations of offshore wind turbine monopile foundations based on large-scale models, to solve the problems of reliance on manual experience, low design efficiency, and insufficient rationality of initial configurations in existing technologies. This scheme uses standardized engineering input templates and target site soil parameters as data foundations. It constructs a constraint set based on specification limits, enterprise design rules, manufacturing and construction restrictions, site-specific conditions, and historical project statistical intervals. Candidate parameters are generated through a constraint-consistent sampling path and a large model generation path. Deterministic geometric relationships are then used to solve for the upper straight section, conical transition section, lower straight section, and optional pile tip thickening section. Unit normalization, strong constraint verification, small-step regression of mud penetration depth, and unified table placement are then used to form an initial configuration of the single-pile foundation that can be used for subsequent structural calculations. This method forms a closed-loop process of "parameter reading—soil layer summary—constraint construction—dual-path generation—geometric solution—strong verification regression—unified table placement."

[0025] According to an embodiment of the present invention, an embodiment of a method for generating the initial configuration of a monopile foundation for offshore wind power is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] This embodiment provides a method for generating the initial configuration of an offshore wind turbine monopile foundation. Figure 1 This is a flowchart of a method for generating the initial configuration of an offshore wind turbine monopile foundation according to an embodiment of the present invention, as shown below. Figure 1 As shown, the process includes the following steps: Step S101: Obtain the engineering parameters and soil parameters of the target location.

[0027] In this step, obtaining the engineering parameters and soil parameters of the target site refers to reading the basic design data and geological information of the site or area from a standardized engineering input template (such as an Excel file, database table, JSON file, or engineering data platform interface).

[0028] In this application, a monopile foundation refers to a foundation form that uses a single large-diameter steel pipe pile as the foundation structure supporting an offshore wind turbine generator. This foundation relies on the pile-soil interaction to provide structural bearing capacity by embedding the pile to a certain depth below the seabed. The target turbine location refers to the specific wind turbine location in the offshore wind farm where a monopile foundation needs to be installed. Each turbine location corresponds to independent design conditions and geological background. Engineering parameters include at least the flange top elevation, mud surface elevation, target mud penetration depth, maximum pile diameter, and diameter-to-thickness ratio range—key information used to define design boundaries and structural constraints. Soil parameters include at least the top and bottom soil layer elevations, soil type, buoyant unit weight, internal friction angle, and undrained shear strength—indicators describing the physical and mechanical properties of each soil layer below the seabed.

[0029] This step is the data entry point for the entire method. It aims to transform scattered and heterogeneous raw engineering data into standardized inputs that can be identified and calculated in subsequent steps. This ensures that the data sources are consistent and traceable throughout the entire process, from constraint construction to parameter generation and geometric solution. At the same time, it avoids problems such as unit errors, coordinate system confusion, or omission of key fields that may be caused by manual data processing, thus laying a data foundation for the reliable generation of the initial configuration of the monopile foundation.

[0030] Step S102: Based on engineering parameters and soil parameters, construct a constraint set to constrain the geometric parameters of the single pile foundation.

[0031] This step transforms the various design conditions, geological information, and engineering rules obtained in step S101 into mathematical constraints that can be directly called upon in subsequent parameter generation steps.

[0032] Specifically, the constraint set includes, but is not limited to: a maximum diameter constraint based on the maximum pile diameter, requiring that the outer diameter of each pile segment does not exceed a preset maximum value; a diameter-to-thickness ratio constraint based on the diameter-to-thickness ratio range, requiring that the ratio of the outer diameter of the pile segment to its wall thickness falls within a preset range; a tapered transition segment slope constraint based on the maximum slope of the tapered transition segment, requiring that the diameter change rate of the tapered transition segment does not exceed the limit allowed by manufacturing and structural continuity; and a minimum segment length constraint based on the minimum segment length, requiring that the length of each pile segment is not less than the lower limit required by the manufacturing, modeling, or interface program.

[0033] The significance of this step lies in transforming the original textual or tabular limits, design rules, manufacturing and construction restrictions, and site-specific conditions into computer-recognizable and executable inequalities or interval conditions. This provides a unified boundary basis for sampling the feasible domain of the first generated path and the standardized processing of the second generated path, ensuring that the subsequently generated candidate parameters always lie within the feasible design space of the project.

[0034] Step S103: Based on the constraint set, generate a candidate parameter set using either the first generation path or the second generation path; wherein, the first generation path is: sampling within the feasible region defined by the constraint set; the second generation path is: generating candidate parameters through a large model based on engineering parameters, soil layer parameters, and the constraint set, and normalizing the candidate parameters generated by the large model so that they fall within the range defined by the constraint set.

[0035] In this step, the first generation path is to sample within the feasible region defined by the constraint set. Specifically, at least one of Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling can be used to generate candidate parameters. This path is mainly used for fast initialization, generation in a model-free environment, or fallback scenarios when a large model is abnormal. The second generation path generates candidate parameters based on engineering parameters, soil layer parameters, and constraint sets through a large model. Specifically, this involves: constructing prompts containing output format requirements based on the engineering parameters, soil layer parameters, and constraint sets; inputting these prompts into the large model; obtaining candidate parameters generated by the large model according to the output format requirements; and performing normalization processing on the candidate parameters generated by the large model to ensure they fall within the range defined by the constraint sets. This normalization process includes converting the candidate parameters to standard units of measurement and trimming parameters exceeding the boundary of the constraint sets to their boundary values. This path is primarily used to generate more detailed, multi-segment configuration drafts that are closer to actual engineering conditions.

[0036] The two generation paths operate under a unified input, unified constraints, unified verification, and unified table placement mechanism. The second generation path is adopted when the preset conditions are met (such as the number of similar samples in the historical project library is lower than the threshold, the site conditions exceed the historical statistical range, or the proportion of candidate solutions that meet the hard constraints after initial sampling is too low). The first generation path is adopted when the preset conditions are not met or the second generation path is abnormal (such as large model interface abnormality, unparseable returned content, missing JSON fields, or multiple failures of strong verification).

[0037] The core significance of this step lies in ensuring that feasible solutions that meet basic constraints can be obtained quickly under any working condition through the sampling path, while generating better candidate parameters by leveraging the interdisciplinary knowledge fusion and complex relationship modeling capabilities of the large model path. The two paths complement and fall back to each other, thereby ensuring the robustness and engineering usability of the solution while giving full play to the advantages of large model inference.

[0038] Step S104: Based on the candidate parameter set, generate the geometric parameters of each pile segment of the single pile foundation using geometric principles, as the initial configuration of the single pile foundation.

[0039] The core of this step lies in transforming the candidate parameter set generated in step S103 into a continuous, manufacturable, and directly usable complete pile segment geometric description for subsequent structural calculations. Specifically, the monopile foundation is divided into several pile segments, and the candidate parameters are automatically synthesized into a complete monopile foundation configuration through deterministic geometric relationships, without manual intervention or repeated trial calculations. This ensures the geometric continuity and physical consistency of the generated results and provides a standardized and directly usable initial input model for subsequent finite element simulations, structural optimization, and engineering verification.

[0040] The initial configuration generation method for offshore wind power monopile foundations provided in this embodiment constructs a constraint set and adopts a dual-path generation mechanism with a first generation path and a second generation path running in parallel. The first generation path can quickly sample within the constraint feasible domain to obtain feasible solutions that satisfy the basic constraints, while the second generation path can automatically model complex soil characteristics and multi-dimensional constraint relationships using a large model to generate candidate parameters that are closer to the actual engineering situation. The two paths can be selected for use or used as a fallback for each other according to preset conditions, thereby effectively eliminating the reliance on repeated iterations based on human experience, significantly improving the design efficiency and rationality of the initial configuration of the monopile foundation, and providing better initial input for subsequent simulation optimization.

[0041] The embodiment provides another method for generating the initial configuration of an offshore wind turbine monopile foundation. Figure 2 This is a flowchart of another method for generating the initial configuration of an offshore wind turbine monopile foundation according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain the engineering parameters and soil parameters of the target location.

[0042] In one optional implementation, the engineering parameters include at least one of the following: flange top elevation, mud surface elevation, target mud penetration depth, maximum pile diameter, and diameter-to-thickness ratio range; the soil layer parameters include at least one of the following: soil layer top elevation, soil layer bottom elevation, soil type, buoyant unit weight, internal friction angle, and undrained shear strength.

[0043] Specifically, the standardized engineering input template is preferably an Excel file, but it is not limited to this; it can also be a database table, a JSON file, a CSV file, or an engineering data platform interface. This template is not simply a wind farm environmental parameter file, but rather a unified engineering data carrier used for generating the initial configuration of monopile foundations, verifying constraints, displaying results, and integrating with subsequent simulations. The data categories, main contents, and uses carried by the template are shown in Table 1.

[0044] Table 1

[0045] The system reads data from the template, including the flange top elevation, average sea level, mud surface elevation, scour-corrected mud surface elevation, target mud penetration depth, maximum pile diameter, diameter-to-thickness ratio range, maximum slope of the conical transition section, maximum length of the conical transition section, minimum section length, and wall thickness range of the target location.

[0046] For ease of calculation, this embodiment uses an elevation system with the Z-axis pointing upwards and the unit being meters. The elevation of the flange top is denoted as z. flange_top The elevation of the mud surface is denoted as z. mud Then the height H above the mud surface above =z flange_top -z mud When considering scour correction, replace the mud surface elevation with the scour-corrected mud surface elevation z. mud,s H above,s =z flange_top -z mud,s .

[0047] The mud surface elevation z in the subsequent formula mud The control mud surface elevation is used for structural calculations; when considering scour corrections, z mud Take the scour correction mud surface elevation z mud,s .

[0048] The system reads the "Soil Parameters" worksheet from the standardized engineering input template. The names, symbols, and meanings of each field in this worksheet are shown in Table 2. The "Soil Parameters" worksheet refers to the worksheet containing soil layer parameters corresponding to the target turbine location or target area in the target wind farm. Its data sources may include engineering geological survey reports, geological profile data, foundation parameter tables compiled by the design unit, or standardized input files that have been reviewed.

[0049] Table 2

[0050] For each soil layer in the soil layer parameters, its thickness h i It can be calculated using the following formula: h i =z top,i -z bottom,i ; Where z top,i z represents the top elevation of the i-th soil layer. bottom,i Let z represent the bottom elevation of the i-th soil layer, and satisfy z top,i >z bottom,i .

[0051] For the soil layers within the target mud-penetration range, the system calculates the overlap thickness between each soil layer and the mud-penetration interval. Let the target mud-penetration interval be [z]. mud -E,z mud], where E represents the target penetration depth, then the effective participation thickness of the i-th soil layer within the penetration range is: h embed,i =max(0,min(z top,i ,z mud )-max(z bottom,i ,z mud -E)).

[0052] Furthermore, let I soft,i The marker indicating whether the i-th soil layer belongs to a weak soil layer is used. The proportion of weak soil layers within the target mud entry interval is then: P soft =Σ(h embed,i ×I soft,i ) / E.

[0053] The system combines multiple soil layer summaries in elevation order into a soil layer summary text for the target location or target area. This summary can be used as input for large model prompts or as a basis for parameter weights and subsurface segment priorities in constrained consistent sampling paths.

[0054] Step S202: Based on engineering parameters and soil layer parameters, construct a constraint set to constrain the geometric parameters of the single pile foundation. The constraint set includes at least one of the following: maximum diameter constraint, diameter-to-thickness ratio constraint, tapered transition section slope constraint, and minimum section length constraint.

[0055] Specifically, the key parameters in this invention are not entirely equivalent to the data obtained from basic information reading and soil layer data parsing. Instead, they are composed of basic elevation parameters, soil mechanical parameters, and engineering constraint parameters. The basic elevation parameters are obtained from template reading, the soil mechanical parameters are obtained from parsing the "Soil Parameters" worksheet, and the engineering constraint parameters are formed by standard limits, enterprise design rules, manufacturing and construction restrictions, and historical project statistical intervals. The names, symbols, sources, and functions of the key parameters involved in this embodiment are shown in Table 3.

[0056] Table 3

[0057] The constraint set is uniformly expressed as mathematical inequalities, equations, or interval conditions, and can be called by the random sampler, large model cue words, and strong validator.

[0058] In this embodiment, the main constraints include at least one of the following: maximum diameter constraint, diameter-to-thickness ratio constraint, tapered transition section slope constraint, tapered transition section length constraint, minimum section length constraint, geometric continuity constraint, mud penetration depth constraint, and positional rule constraint. The determination formulas and processing methods for each constraint condition are shown in Table 4.

[0059] Table 4

[0060] Taking the aspect ratio constraint as an example, this constraint can be converted into a feasible range of wall thickness: D / R max ≤T≤D / R min ; Where D is the outer diameter of the pile segment, T is the wall thickness, and R is the outer diameter of the pile segment. min and R max These are the lower and upper limits of the aspect ratio, respectively.

[0061] If we further consider manufacturing the minimum wall thickness T min and maximum wall thickness T max Then the feasible range of wall thickness can be expressed as: T∈[max(T min D / R max ),min(T max D / R min )).

[0062] If the wall thickness output by the sampling or large model is T0, the corrected wall thickness can be obtained by truncating the callback: T′=min(max(T0,T lower ),T upper ); Where T lower =max(T min D / R max ), T upper =min(T max D / R min ).

[0063] Step S203: Based on the constraint set, generate a candidate parameter set using either the first generation path or the second generation path.

[0064] In this embodiment, the comparison between the constrained consistent sampling path and the large model generation path is shown in Table 5.

[0065] Table 5

[0066] Specifically, step S203 above includes: Step S2031: When the preset conditions are met, the second generation path is adopted.

[0067] In one optional implementation, the preset conditions include, but are not limited to, the following: The number of samples in the historical project database that are similar to the current target site in terms of water depth, mud surface elevation, soil type, target mud penetration depth, or unit level is lower than the preset threshold. The current site conditions exceed the historical sample statistical range; After initial sampling, the proportion of candidate solutions that satisfy all hard constraints is lower than the preset proportion.

[0068] When any of the above conditions are met, the system determines that the historical prior is insufficient or that the feasible region needs to be narrowed down quickly. At this time, the large model is called to generate candidate parameters.

[0069] In one alternative implementation, the second generation path is as follows: based on engineering parameters, soil parameters, and constraint sets, candidate parameters are generated through a large model, and the candidate parameters generated by the large model are normalized so that they fall within the range defined by the constraint sets.

[0070] Specifically, based on engineering parameters, soil layer parameters, and constraint sets, candidate parameters are generated through a large model, including: Step a1: Based on engineering parameters, soil layer parameters, and constraint sets, construct prompt words that include the output format requirements; Step a2: Input the prompt words into the large model; Step a3: Obtain the candidate parameters generated by the large model according to the output format requirements.

[0071] The candidate parameters generated by the large model are normalized to ensure they fall within the range defined by the constraint set, including: Step b1: Convert the candidate parameters generated by the large model into standard units of measurement. Step b2: Trim parameters that exceed the boundary of the constraint set to the boundary values.

[0072] To prevent unit errors, outliers, or inconsistent constraints in the output of large models, this embodiment sets up a deterministic normalization and correction mechanism on the program side. The system uniformly converts the diameter, wall thickness, length, and elevation of the large model output into meters (m), and identifies possible misalignments in centimeters (cm) or millimeters (mm) based on field names, unit labels, and numerical ranges.

[0073] Safety convergence and feasibility filtering include: unit normalization, boundary pruning, relation verification, mud penetration depth adjustment, and multi-candidate screening. Specifically, parameters exceeding the maximum pile diameter, minimum wall thickness, maximum wall thickness, maximum length of the conical transition section, and maximum slope of the conical transition section are truncated or adjusted; candidate schemes that do not meet the requirements for diameter-to-thickness ratio, geometric continuity, positional regularity, and mud penetration depth are locally modified; if the hard constraints are still not met after modification, they are marked as infeasible and resampling or re-generation of the large model is triggered.

[0074] To further explain the specific formulas and functions of each processing step, this embodiment is summarized in Table 6 below.

[0075] Table 6

[0076] Specifically, the historical project database is used to form statistical priors for parameter sampling. Historical priors include statistical information such as pile diameter, wall thickness, pile length, tapered transition section length, tapered transition section location, and segmentation method under similar water depths, similar unit classes, similar soil conditions, and similar mud penetration depths.

[0077] When the number of samples in the historical project database that are similar to the current target site in terms of water depth, mud surface elevation, soil type, target mud penetration depth, or unit class is lower than a preset threshold, or when the current site conditions exceed the statistical range of historical samples, or when the proportion of candidate solutions that satisfy all hard constraints after initial sampling is lower than a preset proportion, the system determines that historical prior knowledge is insufficient or that the feasible region needs to be narrowed down quickly. In the above situations, the system calls the large model, provides meta-information, suggested ranges, or default parameters based on soil layer summary, engineering constraints, and output format requirements, and then a deterministic program performs safe convergence and feasibility filtering.

[0078] To further illustrate the processing methods under different discrimination scenarios, this embodiment is summarized as shown in Table 7.

[0079] Table 7

[0080] The metadata provided by the large model includes: coordinate system description, unit description, flange top elevation, mean sea level elevation, mud surface elevation, target mud penetration depth, constraint version, generation path, prompt word number, model version, and candidate scheme number. The suggested ranges or default values ​​provided by the large model include: upper reference diameter, target diameter near the mud surface, diameter range of the sub-mud section, wall thickness range, length range of the conical transition section, slope range of the conical transition section, number of sub-mud sections, suggested length of each section, whether the pile tip thickening section is enabled, its length, and the amount of thickening.

[0081] Figure 3 This embodiment illustrates the workflow of the large model generation path and the deterministic validator working together. For example... Figure 3 As shown, the system first constructs prompt words based on soil layer summaries, engineering constraints, and output format requirements, and inputs them into the large model generator to obtain candidate segments. Then, the candidate parameters are normalized in units (e.g., converting cm or mm to m), and feasibility is filtered through hard constraint verification (including diameter-to-thickness ratio, slope of tapered transition section, and depth of penetration into mud). If the verification fails, the large model is guided to be regenerated or adjusted through error vector feedback (e.g., error_embed). Finally, the candidate schemes that pass the verification are frozen and output as Excel files, JSON files, and log files.

[0082] Step S2032: When the preset conditions are not met or the second generation path is abnormal, the first generation path is used.

[0083] The first generation path is to sample within the feasible region defined by the constraint set: at least one of Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling is used to sample within the feasible region defined by the constraint set.

[0084] Specifically, this invention provides a constraint-consistent sampling path and a large model generation path. These two paths are not completely independent, but rather two candidate parameter generation methods under a unified input, unified constraints, unified verification, and unified table placement mechanism. The first generation path (i.e., the constraint-consistent sampling path) is used for rapid initialization, generation in a model-less environment, and rollback when the large model is abnormal. The second generation path (i.e., the large model generation path) is used to generate a multi-segment configuration draft based on soil layer summaries and engineering constraints.

[0085] The first generation path is to perform sampling within the feasible region defined by the constraint set. Specifically, the system generates the upper reference diameter D using at least one of the following methods within the feasible region defined by the constraint set: Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling. top Target diameter D near the mud surface mud Candidate wall thickness T, slope s of tapered transition section c Geometric anchor point parameters, such as the target mud penetration depth E and the position parameters of the conical transition section, are included. The sampling results are not the final configuration, but rather the input to the subsequent geometry solver.

[0086] The above sampling process can be represented as: D top ~Ω Dtop D mud ~Ω Dmud , T~Ω T s c ~Ω s ; Where Ω Dtop Ω Dmud Ω T and Ω s These represent the sampling space after being truncated by the constraint set, the historical statistical interval, and the necessary large model suggestion interval, respectively.

[0087] Pre-screening is performed immediately after sampling, including: D top ≤D max D mud ≤D max ; R min ≤D top / T≤R max ; R min ≤D mud / T≤R max; s c ≤s max ; Candidate parameters that pass the pre-screening are used in subsequent geometric solution steps, while candidate parameters that do not pass the pre-screening are discarded and resampled.

[0088] Step S204: Based on the candidate parameter set, generate the geometric parameters of each pile segment of the single pile foundation using geometric principles, as the initial configuration of the single pile foundation.

[0089] In this embodiment, the single pile foundation has a three-section configuration, see [link / reference]. Figure 4 It includes an upper straight section U, a tapered transition section C, and a lower straight section L. Optionally, a pile end thickening section toe is added at the bottom of the lower straight section.

[0090] Figure 4 The three-section monopile foundation configuration and main geometric parameters in this embodiment are shown. Figure 4 As shown, a single pile foundation consists of an upper straight section, a conical transition section (i.e., a conical section), and a lower straight section from top to bottom. The conical transition section connects the upper straight section and the lower straight section, achieving a smooth transition in pile diameter. Figure 4 The flange top elevation z is marked in the middle. flange_top Mud surface elevation z mud Mean sea level elevation z msl Equal base elevation, and the length L of the upper straight section u Length L of the tapered transition section c Length L of the lower straight section l The slope s of the conical transition section c Key geometric parameters include the positional relationship of the conical transition section and the target mud penetration depth E. Optionally, a thickened pile end section toe can be added to the bottom of the lower straight section, with a wall thickness of T equal to the wall thickness of the lower straight section plus the thickening amount ΔT.

[0091] Each pile segment can be uniformly represented as: S i ={z top,i ,L i D top,i D bottom,i ,T i ,type i}; Where z top,i Let L be the top elevation of the i-th segment. i D is the length of this segment. top,i and D bottom,i These are the outer diameters of the top and bottom of the segment, respectively, T i For wall thickness, type i It is a segment type.

[0092] The elevation of the bottom of segment i is: z bottom,i=z top,i -L i .

[0093] Adjacent segments should satisfy geometric continuity constraints: z top,i+1 =z bottom,i .

[0094] If the i-th segment is a cylindrical segment, then D should be satisfied. top,i =D bottom,i If the i-th segment is a conical transition segment, then its actual slope is: s i =|D top,i -D bottom,i | / L i ; And should satisfy s i ≤s max , where s max This is the preset maximum slope.

[0095] Given the outer diameter D at the top of the tapered transition section top,c The outer diameter D at the bottom of the tapered transition section bottom,c and the slope s of the conical transition section c The theoretical length of the tapered transition section is: L c =|D top,c -D bottom,c | / s c .

[0096] If L c Exceeding the maximum length L of the tapered transition section c,max Then adjust the end diameter and slope of the tapered transition section, or regenerate the anchor point parameters, so that the adjusted tapered transition section simultaneously satisfies both length and slope constraints: L′ c ≤L c,max , and |D top,c -D bottom,c | / L′ c ≤s max .

[0097] Specifically, step S204 above includes: Step S2041: Calculate the length, top elevation, and bottom elevation of the conical transition section based on the conical transition section parameters in the candidate parameter set. Step S2042: Calculate the length of the upper straight section based on the top elevation of the flange and the top elevation of the tapered transition section; Step S2043: Calculate the length of the lower straight section based on the bottom elevation, mud surface elevation, and target mud penetration depth of the conical transition section.

[0098] Specifically, the tapered transition section should be located above the mud surface. First, determine the total length L of the tapered transition section. c Combined with the mud surface elevation z mud Determine the top and bottom elevations of the conical transition section to satisfy the following conditions: z bottom,c ≥z mud ; L c,below =0; Among them, z mud This represents the mud surface elevation; when considering scour correction, the scour-corrected mud surface elevation is used; L c,below This represents the length of the conical transition section below the mud surface, and its value is 0.

[0099] The bottom and top elevations of the conical transition section can be expressed as: z bottom,c ≥z mud ; z top,c =z bottom,c +L c .

[0100] If the bottom elevation of the conical transition section is z bottom,c If the elevation is below the mud surface elevation, the position constraint is deemed unacceptable and the system is regenerated.

[0101] In one embodiment, the bottom elevation of the tapered transition section can be specified by engineering rules within the range that satisfies the above-mentioned positional constraints, or the bottom elevation of the tapered transition section can be automatically determined according to structural layout rules: z bottom,c ≥z mud ; z top,c =z bottom,c +L c ; This method ensures that the conical transition section is located above the mud surface.

[0102] Once the top and bottom elevations of the conical transition section are determined, the length L of the upper straight section is... u and the length L of the lower straight section l The solution can be obtained using the following formula: L u =z flange_top -z top,c ; L l =(z bottom,c -z mud )+E; Where L u L represents the length of the upper straight section. l z represents the length of the lower straight segment. bottom,c -z mudE represents the length of the lower straight section above the mud surface, and E represents the effective depth of penetration into the mud below the mud surface.

[0103] The system further checks L u and L l Does the minimum segment length constraint hold? L u ≥L min L l ≥L min .

[0104] If the length of a certain segment is less than L min If necessary, adjust the position of the conical transition section, recalculate the bottom elevation of the conical transition section, or resample the geometric anchor point parameters.

[0105] In an optional implementation, the system may add a thickened pile end section (toe) at the bottom of the lower straight section, the data structure of which can be represented as follows: S toe ={z top,toe ,L toe D toe D toe ,T+ΔT,type=toe}.

[0106] This section is used to improve the local stiffness and resistance to local deformation of the pile tip, and reduce the risk of edge curling, ellipticization, local buckling or end damage of the pile tip during pile driving, installation or other processes.

[0107] When the target penetration depth E is defined as the effective penetration control length excluding the pile tip thickening section, the length of the pile tip thickening section can be treated as an additional length and is not included in the target penetration depth E, but is included in the total physical length of the single pile; when the target penetration depth E is defined as the total penetration length from the mud surface to the pile tip, the pile tip thickening section is included in the target penetration depth E. Under the caliber of the effective penetration control length excluding the pile tip thickening section: L total =L u +L c +L l +L toe ; L embed =L l -(z bottom,c -z mud )=E; Where L total L represents the total physical length of a single pile. embed L represents the effective depth of mud penetration used for mud penetration depth verification. toe This indicates the length of the thickened section at the pile end.

[0108] In another optional implementation, after generating the geometric parameters of each pile segment of the single pile foundation, the method of this embodiment further includes a mud penetration depth callback step, specifically: Step c1: Calculate the cumulative length of the pile segment below the mud.

[0109] For any pile segment, its effective length L under the mud below,i It can be automatically calculated from the intersection relationship between the pile segment and the mud surface elevation: L below,i =max(0,min(z top,i ,z mud )-z bottom,i ).

[0110] When a thickened section exists at the pile tip and this section is not included in the target penetration depth, the cumulative subsurface length excluding the thickened section is: L below , no_toe =Σ(L below,i ×I i ); Where I i To be included in the calculation. When the target mud penetration depth is included in the calculation for the i-th segment, I i =1, when the i-th segment is a thickened segment at the pile tip or other additional segment not included in the target mud penetration depth. i =0.

[0111] Step c2: Calculate the deviation between the cumulative mud depth and the target mud penetration depth.

[0112] The deviation between the two is used to verify the consistency of the mud penetration depth, specifically as follows: |EL below,no_toe |≤tol embed ; Where E is the target depth into the mud, and tol embed This is the preset tolerance.

[0113] Step c3: When the absolute value of the deviation exceeds the preset tolerance, adjust the length of the lower straight section and regenerate the geometric parameters.

[0114] If the above conditions are not met, the system will prioritize performing small-step pullbacks on the cylindrical sections below the mud surface and not thickened at the pile tip. Let the deviation ΔE = EL. below,no_toe Then the k-th callback can be represented as: L l (k+1) =L l (k) +sign(ΔE)×step k ; Step k =min(step max,max(step min ,α×|ΔE|)), step k This represents the step size for the k-th callback, which can be an adaptive step size ranging from 0.05m to 0.50m. α represents the scaling factor. After each callback, the system recalculates the elevation of the downstream segment and performs a hard constraint check.

[0115] In this embodiment, a unified strong verification is also performed on the output results of the randomly generated path and the large model generated path. Figure 5 This is a flowchart illustrating the strong verification and error vector feedback process in an embodiment of the present invention. Figure 5 As shown, after receiving candidate parameters, the system sequentially performs hard boundary checks (including the maximum diameter constraint D≤D). max Wall thickness manufacturing constraints T min ≤T≤T max ), consistency verification of relationships (including aspect ratio constraint R) min ≤D / T≤R max The slope constraint of the tapered transition section is slope ≤ s max ), and location and mud insertion depth verification (including mud insertion depth consistency | EL) below |≤tol、Geometric continuity constraint z top,i+1 =z bottom,i The validation results are output in the format of "pass / fail + reason location": for solutions that pass the validation, a table insertion operation is performed; for solutions that fail, an error vector (e.g., error...) is generated. D =DD max error R =max(R min -D / T,D / TR max ), error E =EL below,no_toe (etc.), and trigger a callback or regeneration mechanism.

[0116] The verification content includes: unit consistency, geometric continuity, cylindrical section diameter consistency, conical transition section slope, conical transition section length, maximum diameter, diameter-to-thickness ratio, minimum section length, mud penetration depth, position regularity, and pile tip thickening section regularity. The verification results are output in the form of "pass / fail + reason location".

[0117] Once a candidate scheme passes strong validation, the system automatically copies the input template and generates a new scheme file. The pile segments above the mud surface are written to the "Pile Information Above Mud Surface" worksheet, and the cylindrical segments below the mud surface are written to the "Pile Dimensions Below Mud Surface" worksheet. If the target calculation program requires diameter and wall thickness to be expressed in centimeters (cm), the system automatically performs unit conversion. D cm =100×D m ;Tcm =100×T m ; Where D cm and T cm D represents the pile segment diameter and wall thickness, respectively, in cm. m and T m These represent the pile segment diameter and wall thickness, respectively, in meters (m).

[0118] The effective length above the mud surface can be expressed as: L above,i =max(0,z top,i -max(z bottom,i ,z mud )).

[0119] The effective length below the mud surface can be expressed as: L below,i =max(0,min(z top,i ,z mud )-z bottom,i ).

[0120] Figure 6 This is a schematic diagram illustrating the unified table placement and simulation interface connection in an embodiment of the present invention. For example... Figure 6 As shown, the system takes an input template (containing data such as elevation, soil parameters, and constraints) and the generated results (segments JSON) as input. It performs unit conversion (such as converting diameter and wall thickness to cm and retaining length as m) and effective length calculation through a field mapper, generating a pile information table above the mud surface (containing top elevation, outer diameter, wall thickness, and effective length) and a pile dimension table below the mud surface (containing diameter, wall thickness, and length). At the same time, it outputs a verification list and log file (containing pass / fail status and reason location), which are used to connect with subsequent structural calculation programs and engineer review.

[0121] In a preferred embodiment, the system can perform a lightweight evaluation of multiple candidate solutions that pass the hard constraints. This evaluation does not replace subsequent finite element calculations, but is only used to select a solution that is relatively suitable for entering the simulation from multiple initial feasible configurations.

[0122] The scoring function is expressed as: Score=w1·P constraint +w2·P smooth +w3·P discrete +w4·P soil +w5·P volume ; Where P constraint P represents the constraint margin score. smooth P represents the geometric smoothness score. discrete P represents the manufacturing discretization score.soil P represents the soil compatibility score. volume The rating represents the amount of work done on behalf of the agent, with w1 to w5 representing the corresponding weights.

[0123] For the i-th cylindrical segment, its steel pipe volume can be represented as: V i =π / 4×[D² i -(D i -2T i )²]×L i .

[0124] For tapered transition sections, the volume can be approximated by using the average diameter of the upper and lower ends: D c =(D top,c +D bottom,c ) / 2; V c ≈π / 4×[D² c -(D c -2T c )²]×L c .

[0125] It should be noted that the engineering quantity proxy is only used for ranking among candidate initial configurations. The final configuration should still be based on subsequent structural calculations, specification verification, and engineering review.

[0126] The initial configuration generation method for offshore wind power monopile foundations provided in this embodiment, compared with the traditional method of gradually setting initial parameters based on manual experience, can form a traceable, verifiable, and documented initial configuration generation process based on standardized engineering input templates and target site soil parameters.

[0127] First, this invention uses a standardized engineering input template to uniformly read flange top elevation, average sea level elevation, mud surface elevation, target mud penetration depth, soil layer parameters, and structural constraint parameters. This helps reduce repetitive work of manually compiling input data and reduces input errors caused by inconsistencies in units, coordinate systems, or table formats.

[0128] Secondly, this invention constructs a constraint set by standard limits, enterprise design rules, manufacturing and construction restrictions, site conditions and historical project statistical intervals, so that the randomly generated path is no longer an unconstrained random selection, but a constrained consistent sampling within the engineering feasible domain, and automatically derives the three-segment configuration through geometric formulas, thereby reducing the probability of invalid solutions entering subsequent simulations.

[0129] Furthermore, this invention limits the output of the large model to structured data containing meta-information, constraint mirrors, and pile segment arrays. It also sets up unit normalization, outlier identification, boundary trimming, diameter-to-thickness ratio verification, tapered transition section verification, geometric continuity verification, and mud penetration depth callback mechanisms on the program side, which helps to solve the uncontrollable problems that may exist when the large model directly generates engineering parameters.

[0130] Furthermore, this invention clarifies the length, elevation, and diameter relationships between the upper straight section, the conical transition section, the lower straight section, and the optional pile end thickening section through a unified coordinate system and formulaic geometric solution. This enables the generated results to be expressed in a unified segments data structure, facilitating subsequent writing into engineering templates and integration into structural calculation software.

[0131] Finally, this invention includes mechanisms for large model anomaly rollback, sampling regeneration, and small step-size callback. When the large model interface malfunctions, the output format is incomplete, or strong validation fails, the system can roll back to the constraint-consistent sampling path, outputting initial candidate solutions that can be reviewed by engineers or used for subsequent simulations, thereby enhancing the robustness of the method under different operating environments.

[0132] In summary, this invention can improve the automation, constraint compliance, engineering adaptability, and result traceability of the initial configuration generation process for monopile foundations without pre-setting unverified quantitative effects, providing a more standardized initial input for subsequent finite element simulation, structural optimization, and engineer review.

[0133] This embodiment also provides another method for generating the initial configuration of offshore wind turbine monopile foundations, applicable to the automatic generation of the initial configuration of offshore wind turbine monopile foundations before scheme design, preliminary design, batch turbine site selection, finite element simulation preprocessing, and structural optimization. This method uses an engineering input template of the target turbine site or target area as a basis, and through reading basic elevation parameters, analyzing soil layer parameters, constructing constraint sets, performing dual-path generation, geometric solving, and strong verification, finally outputs a new scheme file that can be called by subsequent structural calculation programs.

[0134] (1) Basic parameter reading and engineering input preprocessing.

[0135] a. The system receives the standardized engineering input template and target location number specified by the user. The template can be an Excel file and contains basic elevation information, soil parameter worksheets, structural constraint areas, pile information tables above the mud surface, and pile dimension tables below the mud surface.

[0136] b. The system reads fields such as flange top elevation, average sea level elevation, mud surface elevation, scour-corrected mud surface elevation, target mud penetration depth, maximum pile diameter, diameter-to-thickness ratio range, maximum slope of tapered transition section, maximum length of tapered transition section, minimum section length, and wall thickness range.

[0137] c. The system performs unit checks and missing field checks on the read results. If a field is missing, it is filled in according to the standard limits, enterprise rules, historical project statistical ranges, or large model suggestions, and the source and version of the fill are recorded.

[0138] (2) Soil layer data analysis.

[0139] a. The system reads the "Soil Parameters" worksheet and extracts the soil layer number, soil layer name, soil type, top elevation of soil layer, bottom elevation of soil layer, buoyant unit weight, internal friction angle, undrained shear strength, side friction resistance, end resistance, and strain parameters of the target location or target area.

[0140] b. The system sorts the soil layers from top to bottom according to elevation and calculates the thickness of each layer. For soil layers within the target mud entry range, the system calculates their effective mud entry thickness, forming a summary of the soil layer distribution within the target mud entry range.

[0141] c. The system converts soil layer summaries into two formats: text and structured data. The text format is used for inputting prompts for large models, while the structured data format is used for sampling weights, sub-gravel segmentation, and subsequent validation.

[0142] (3) Constraint set generation.

[0143] a. The system generates maximum diameter constraints, diameter-to-thickness ratio constraints, tapered transition section slope constraints, tapered transition section length constraints, minimum section length constraints, mud penetration depth constraints, and geometric continuity constraints based on template fields and built-in rules.

[0144] b. The system generates positional constraints based on the mean sea level elevation, mud surface elevation, and structural layout rules. For example, the bottom of the upper straight section must meet a specified positional requirement relative to the mean sea level elevation, and the conical transition section must meet the rule of being located above the mud surface.

[0145] c. The system expresses the above constraints as a set of constraints that can be used by the sampler, large model cue words, and strong validator.

[0146] (4) Parameter generation path selection.

[0147] a. When the user selects random generation or the large model interface is unavailable, the system enters the constraint-consistent sampling path. The system samples geometric anchor point parameters such as the upper reference diameter, the target diameter near the mud surface, the candidate wall thickness, the slope of the tapered transition section, and the position parameters of the tapered transition section within the feasible region defined by the constraint set.

[0148] b. When the user selects large model generation and the large model interface is normal, the system constructs a prompt containing basic elevation parameters, soil layer summary, engineering constraints and output JSONSchema, and calls the large model to generate multi-segment candidate pile segment data.

[0149] c. Regardless of the path used, the output results are converted into a unified segments data structure and then enter the subsequent geometric solution and strong verification process.

[0150] (5) Solving the three-segment configuration.

[0151] a. For constrained consistent sampling paths, the system follows D top D mud and the slope s of the conical transition section c Calculate the length L of the tapered transition section c And check L c Does it exceed the maximum length of the tapered transition section?

[0152] b. The system determines the position of the conical transition section according to engineering rules, so that the conical transition section is above the mud surface; if its bottom elevation is lower than the mud surface elevation, it triggers regeneration.

[0153] c. The system calculates the length of the upper and lower straight sections based on the top and bottom elevations of the tapered transition section, and calculates or adjusts the wall thickness of each section according to the diameter-to-thickness ratio constraint.

[0154] (6) Normalization and correction of large model results.

[0155] a. For each pile segment output from the large model, the system checks whether the segments such as segment number, type, top elevation, length, top diameter, bottom diameter, wall thickness, location label, and whether it is a TOE segment are complete.

[0156] b. The system converts all lengths, elevations, diameters, and wall thicknesses to meters and identifies possible misalignments in mm or cm.

[0157] c. The system performs boundary trimming, relation verification, and necessary callbacks for the outer diameter, wall thickness, tapered transition section length, tapered transition section slope, and mud penetration depth. If the hard constraints are still not met after the callback, the error vector is fed back to the large model for regeneration, or the system reverts to the constraint-consistent sampling path.

[0158] (7) Strong verification.

[0159] a. The system checks the geometric continuity segment by segment, that is, the bottom elevation of each segment should be equal to the top elevation of the next segment.

[0160] b. The system checks whether the top and bottom diameters of the cylindrical section are consistent, whether the slope of the tapered transition section is less than or equal to the maximum allowable slope, and whether the length of the tapered transition section is less than or equal to the maximum allowable length.

[0161] c. The system checks whether all pile segments meet the constraints of maximum diameter, diameter-to-thickness ratio, minimum segment length, mud penetration depth, and position rules. The verification results are output as an itemized list.

[0162] (8) Small step-by-step pullback in mud depth.

[0163] a. If the difference between the cumulative mud depth and the target mud penetration depth exceeds the tolerance, the system calculates the deviation ΔE.

[0164] b. When the actual mud penetration length is insufficient, the system prioritizes increasing the length of the cylindrical section below the mud surface and not thickened at the pile end; when the actual mud penetration length is too long, the system prioritizes reducing the length of that section.

[0165] c. After each adjustment, the system cascades and updates the elevation of the downstream segment, and re-executes checks on geometric continuity, diameter-to-thickness ratio, tapered transition segment constraints, minimum segment length, and mud penetration depth. If a callback causes hard constraints to be broken, the adjustment is rolled back and the step size is reduced or the system is regenerated.

[0166] (9) Treatment of thickened section at pile end.

[0167] a. If the template or generated result specifies that the pile end thickening section is enabled, the system will add a toe section at the bottom of the lower straight section.

[0168] b. The outer diameter of the toe section is usually the same as that of the lower straight section, and the wall thickness is the wall thickness of the lower straight section plus the thickening amount ΔT.

[0169] c. The toe segment is included in the total physical length of a single pile, but not in the target penetration depth E. The system still performs manufacturing feasibility, upper limit of wall thickness, and geometric continuity checks on the toe segment.

[0170] (10) Unify table and file output.

[0171] a. The system copies the original input template and generates a new solution file.

[0172] b. The system writes the top elevation, outer diameter, wall thickness, effective length, and diameters at both ends of the tapered transition section of the pile segment above the mud surface into the "Information on Pile Above Mud Surface" worksheet.

[0173] c. The system records the diameter, wall thickness, and effective length of the cylindrical section below the mud surface into the "Dimensions of Piles Below Mud Surface" worksheet. The diameter and wall thickness can be converted to cm as required by subsequent simulation software, while the length is retained in meters.

[0174] d. The system synchronously outputs the segmentsJSON file, a strong validation list, and generation logs to support engineer review, solution traceability, and subsequent optimization.

[0175] (11) Reasoning anomalies and rollback mechanisms.

[0176] a. When the large model interface is abnormal, the returned content cannot be parsed, JSON fields are missing, or strong validation fails multiple times, the system will automatically fall back to the constraint consistency sampling path.

[0177] b. When the sampling path continuously generates infeasible solutions, the system shrinks the sampling range, adjusts the prior weights, or prompts the user to check the key fields in the input template.

[0178] c. The above-mentioned rollback mechanism can ensure that initial candidate schemes that can be manually reviewed or structurally calculated are output as much as possible under different operating environments.

[0179] (12) Engineering verification methods.

[0180] a. In practical applications, representative target locations or areas can be selected, and their standardized engineering input templates and soil parameters can be read to generate the initial configuration of a single pile foundation.

[0181] b. After the generated results are written into the engineering template, they can be connected to the subsequent structural calculation process to analyze static, modal, fatigue or seismic conditions.

[0182] c. By comparing the verification checklists, finite element calculation input files, and engineer review comments before and after generation, the improvement effect of this method on the efficiency of initial configuration generation, constraint compliance, and subsequent adjustment workload can be evaluated. This verification process does not presuppose unproven quantitative effects; the specific effects are based on actual engineering data and calculation results.

[0183] This embodiment also provides an initial configuration generation device for offshore wind turbine monopile foundations. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0184] This embodiment provides a device for generating the initial configuration of a monopile foundation for offshore wind power, such as... Figure 7 As shown, it includes: The data acquisition module 701 is used to acquire the engineering parameters and soil parameters of the target location; Constraint construction module 702 is used to construct a set of constraints for constraining the geometric parameters of a single pile foundation based on engineering parameters and soil parameters; The dual-path generation module 703 is used to generate a set of candidate parameters based on a set of constraints, using either a first generation path or a second generation path. The first generation path involves sampling within the feasible region defined by the set of constraints. The second generation path involves generating candidate parameters from a large model based on engineering parameters, soil parameters, and the set of constraints, and then normalizing the candidate parameters generated from the large model to ensure they fall within the range defined by the set of constraints. The geometry solver module 704 is used to generate the geometric parameters of each pile segment of a monopile foundation based on the candidate parameter set and geometric principles, as the initial configuration of the monopile foundation.

[0185] In some alternative implementations, the dual-path generation module 703 includes: The second generation path execution unit is used to adopt the second generation path when preset conditions are met. The second generation path is as follows: based on engineering parameters, soil layer parameters, and constraint sets, candidate parameters are generated through a large model, and the candidate parameters generated by the large model are normalized to ensure that they fall within the range defined by the constraint sets.

[0186] The first generation path execution unit is used to adopt the first generation path when the preset conditions are not met or the second generation path is abnormal. The first generation path is: sampling within the feasible region defined by the constraint set: sampling is performed within the feasible region defined by the constraint set using at least one of Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling.

[0187] In some alternative implementations, the geometry solving module 704 includes: The tapered transition section solution unit is used to calculate the length, top elevation, and bottom elevation of the tapered transition section based on the tapered transition section parameters in the candidate parameter set. The upper straight section solution unit is used to calculate the length of the upper straight section based on the top elevation of the flange and the top elevation of the tapered transition section; The lower straight section solution unit is used to calculate the length of the lower straight section based on the bottom elevation of the conical transition section, the mud surface elevation, and the target mud penetration depth.

[0188] In another optional embodiment, the device of this embodiment further includes a mud penetration depth callback module, which is used to: calculate the cumulative mud penetration length of each pile segment; calculate the deviation between the cumulative mud penetration length and the target mud penetration depth; and when the absolute value of the deviation exceeds the preset tolerance, adjust the length of the lower straight section and regenerate the geometric parameters.

[0189] The offshore wind power monopile foundation initial configuration generation device provided in this embodiment of the invention can execute the offshore wind power monopile foundation initial configuration generation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0190] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0191] The following is a detailed reference. Figure 8This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 801, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 802 or a program loaded from memory 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the electronic device. The processor 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0192] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 808 including, for example, magnetic tapes, hard disks, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0193] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 809, or installed from a memory 808, or installed from a ROM 802. When the computer program is executed by the processor 801, it performs the functions defined in the method for generating the initial configuration of an offshore wind turbine monopile foundation according to embodiments of the present invention.

[0194] Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0195] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the method for generating the initial configuration of an offshore wind turbine monopile foundation shown in the above embodiments is implemented.

[0196] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0197] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for generating the initial configuration of a monopile foundation for offshore wind power, characterized in that, The method includes: Obtain the engineering parameters and soil parameters of the target location; Based on the aforementioned engineering parameters and soil layer parameters, a constraint set is constructed to constrain the geometric parameters of a single pile foundation. Based on the constraint set, a candidate parameter set is generated using either a first generation path or a second generation path. The second generation path is used when a preset condition is met; the first generation path is used when the preset condition is not met or the second generation path is abnormal. The first generation path involves sampling within the feasible region defined by the constraint set. The second generation path involves generating candidate parameters through a large model based on the engineering parameters, soil parameters, and constraint set, and then normalizing the candidate parameters generated by the large model to ensure they fall within the range defined by the constraint set. Based on the candidate parameter set, the geometric parameters of each pile segment of the monopile foundation are generated using geometric principles, serving as the initial configuration of the monopile foundation.

2. The method for generating the initial configuration of an offshore wind turbine monopile foundation according to claim 1, characterized in that, The engineering parameters include at least one of the following: flange top elevation, mud surface elevation, average sea level elevation, target mud penetration depth, maximum pile diameter, and diameter-to-thickness ratio range; The soil layer parameters include at least one of the following: top elevation of soil layer, bottom elevation of soil layer, soil type, buoyant unit weight, internal friction angle, and undrained shear strength; The constraint set includes at least one of the following: maximum diameter constraint, diameter-to-thickness ratio constraint, tapered transition section slope constraint, and minimum section length constraint.

3. The method for generating the initial configuration of an offshore wind turbine monopile foundation according to claim 1, characterized in that, The sampling within the feasible region defined by the constraint set includes: Sampling is performed within the feasible region defined by the constraint set using at least one of Latin hypercube sampling, Sobol quasi-random sampling, or truncated distribution sampling.

4. The method for generating the initial configuration of an offshore wind turbine monopile foundation according to claim 1, characterized in that, The process of generating candidate parameters based on the engineering parameters, soil layer parameters, and constraint set through a large model includes: Based on the aforementioned engineering parameters, soil layer parameters, and constraint set, a prompt word containing output format requirements is constructed; Input the prompt words into the large model; Obtain the candidate parameters generated by the large model according to the output format requirements.

5. The method for generating the initial configuration of an offshore wind turbine monopile foundation according to claim 1, characterized in that, The normalization process for the candidate parameters generated by the large model, ensuring they fall within the range defined by the constraint set, includes: Convert the candidate parameters generated by the large model into standard units of measurement. Parameters that exceed the boundary of the constraint set are clipped to the boundary value.

6. The method for generating the initial configuration of an offshore wind turbine monopile foundation according to claim 2, characterized in that, The single pile foundation includes an upper straight section, a conical transition section, and a lower straight section, wherein the conical transition section is located above the mud surface; The generation of geometric parameters for each pile segment of a monopile foundation using geometric principles includes: Calculate the length, top elevation, and bottom elevation of the tapered transition section based on the tapered transition section parameters in the candidate parameter set; Calculate the length of the upper straight section based on the top elevation of the flange and the top elevation of the tapered transition section; The length of the lower straight section is calculated based on the bottom elevation, mud surface elevation, and target mud penetration depth of the conical transition section.

7. The method for generating the initial configuration of an offshore wind turbine monopile foundation according to claim 6, characterized in that, After generating the geometric parameters of each pile segment of the monopile foundation, the process also includes: Calculate the cumulative length of the pile segment below the mud. Calculate the deviation between the cumulative mud depth and the target mud penetration depth; When the absolute value of the deviation exceeds the preset tolerance, the length of the lower straight section is adjusted and the geometric parameters are regenerated.

8. A device for generating the initial configuration of a monopile foundation for offshore wind power, characterized in that, The device includes: The data acquisition module is used to acquire the engineering parameters and soil parameters of the target location; The constraint construction module is used to construct a constraint set for constraining the geometric parameters of a single pile foundation based on the engineering parameters and soil layer parameters. A dual-path generation module is used to generate a candidate parameter set based on the constraint set, using either a first generation path or a second generation path. The second generation path is used when preset conditions are met; the first generation path is used when the preset conditions are not met or the second generation path is abnormal. The first generation path involves sampling within the feasible region defined by the constraint set. The second generation path involves generating candidate parameters through a large model based on the engineering parameters, soil layer parameters, and the constraint set, and then normalizing the candidate parameters generated by the large model to ensure they fall within the range defined by the constraint set. The geometry solving module is used to generate the geometric parameters of each pile segment of the monopile foundation based on the candidate parameter set and using geometric principles, as the initial configuration of the monopile foundation.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the method for generating the initial configuration of an offshore wind power monopile foundation as described in any one of claims 1 to 7.

Images