A tension leg type floating wind turbine multi-parameter automatic collaborative optimization design method

Abstract

The present application belongs to the field of marine renewable energy and floating structure engineering optimization design, and discloses a tension leg type floating wind turbine multi-parameter automatic collaborative optimization design method and system. The method realizes the integrated efficient optimization of the floating body and the anchoring system by fusing parameterized geometric modeling, open source frequency domain hydrodynamic calculation, fast load evaluation and genetic algorithm. The present application realizes efficient automatic design, avoids the unstable interface problem, and the process is reproducible and expandable. The present application can optimize in a larger design space, obtain a comprehensive design scheme with better performance matching of the floating body and the anchoring system and lower cost of the whole system, effectively balance efficiency and reliability, and stably handle complex constraints based on the genetic algorithm adaptive optimization mechanism of multi-constraint punishment.

Description

Technical Field

[0001] This invention belongs to, but is not limited to, the field of marine renewable energy and floating structure engineering optimization design technology, and particularly relates to a multi-parameter automated collaborative optimization design method and system for tension leg floating wind turbines. Background Technology

[0002] As offshore wind power expands into deeper waters, floating wind power has become a core technology direction. Among them, tension leg platforms are particularly suitable for supporting high-power wind turbines due to their excellent motion stability and small horizontal displacement amplitude. However, the design of tension leg floating wind turbines is a typical complex engineering optimization problem with multiple variables, multiple constraints, and strong coupling, involving the coordination of multiple dimensions such as the main dimensions of the floating body, ballast configuration, and mooring stiffness.

[0003] Currently, the design optimization of tension leg type floating wind turbines still faces the following technical bottlenecks:

[0004] (1) Fragmented design process and low degree of automation: Traditional design relies on manual iteration. The floating body design, mooring design, hydrodynamic analysis, cost estimation and other links are independent of each other and rely on commercial software interfaces, resulting in long design cycle, low efficiency and difficulty in achieving systematic exploration of large-scale parameter space.

[0005] (2) Limited optimization dimensions and insufficient synergy: Existing studies mostly focus on optimizing local dimensions of the float (such as the radius of the heave plate, the diameter of the column, etc., usually about 3), and fail to integrate and synergistically optimize the mooring system (tendon length, diameter, pretension) as an endogenous variable, ignoring the float-mooring coupling effect, making it difficult to obtain the global optimal solution.

[0006] (3) It is difficult to balance computational fidelity and efficiency: In the conceptual design stage, if high-fidelity time-domain coupled simulation is used directly for evaluation, although the results are accurate, the computational cost is too high and cannot support large-scale parameter screening; if the constraints are not considered enough, the design scheme may fail in the detailed design stage.

[0007] In existing tension leg floating wind turbine platforms, the engineering design and industrial application processes generally employ a phased, sequential design workflow dominated by manual experience. This results in fragmented analytical models across different disciplines, lacking a unified collaborative mechanism between geometric design, hydrodynamic analysis, motion response verification, and cost assessment. When the platform encounters issues such as periodic avoidance, average displacement, or tendon stress that fail to meet requirements during subsequent verification stages, designers must manually backtrack and adjust structural parameters repeatedly. This leads to low design efficiency, high iteration costs, and difficulty in obtaining a globally optimal solution under complex constraints. Especially with the ever-increasing size of large-capacity wind turbine platforms, the aforementioned design model struggles to simultaneously consider dynamic safety and engineering economy, becoming a core technological bottleneck restricting the large-scale application of tension leg floating wind turbine platforms. Summary of the Invention

[0008] This invention provides a collaborative design optimization method that enables automatic linkage of multidisciplinary performance during the design phase and directly leads to the cost-optimal solution while satisfying key dynamic and structural constraints. This overcomes the problems of existing technologies that rely on human experience, have low design efficiency, and produce unstable optimization results.

[0009] To address the aforementioned technical problems, this invention proposes a multi-parameter automated collaborative optimization design method for tension leg floating wind turbines. This method integrates geometric modeling, hydrostatic and mass calculations, frequency domain hydrodynamic analysis, motion response evaluation, and cost estimation by constructing an automated design-evaluation-optimization closed loop centered on parametric design vectors. In this method, platform structural parameters are systematically introduced into the design space as independent design variables. Any design scheme automatically generates a geometric model through parametric modeling and completes ballast and tendon pretension matching under static equilibrium conditions. Subsequently, based on frequency domain hydrodynamic calculations and system motion characteristics, key dynamic indicators such as the platform's average drift displacement and natural period are obtained. On this basis, an optimization model is constructed with the estimated cost as the objective and periodic avoidance, displacement limitation, and structural safety as constraints. A genetic algorithm is then used to adaptively optimize within the entire design space, outputting the optimal design scheme that meets the engineering constraints.

[0010] Furthermore, parametric geometric modeling and automatic mesh generation mechanisms, fast average drift evaluation methods, composite penalty fitness functions, and adaptive convergence strategies are introduced to enable the optimization process to converge stably while ensuring computational efficiency. Through the synergistic effect of multi-layer constraints, the design results are guided to achieve a balance between dynamic safety and engineering economy.

[0011] This invention provides an automated multidisciplinary performance evaluation method for collaborative design optimization of tension leg type floating wind turbine platforms. It automatically outputs evaluation results for constraint determination and cost calculation for any given design scheme. The evaluation method includes:

[0012] E1 automatically generates the platform geometry model and generates surface mesh;

[0013] E2 automatically calculates hydrostatic and mass characteristics and determines the ballast in conjunction with the static balance and satisfies the set tendon pretension.

[0014] E3 automatically calculates frequency domain hydrodynamic coefficients and obtains wave excitation;

[0015] E4, quickly assess the average drift displacement;

[0016] E5, the natural period is solved based on the platform mass characteristics, hydrodynamic added mass, hydrostatic recovery characteristics and tendon linearized stiffness;

[0017] E6 outputs a vector of evaluation results including average drift displacement, natural period, tendon tension and strength criteria, displacement, and cost, which can be used by the optimization module.

[0018] Furthermore, the average sway displacement in E4 is estimated by dividing the sum of the average wind thrust and the average wave drift force by the total stiffness in the sway direction. The average wind thrust is obtained by interpolation from the curves of rated wind speed and wind turbine thrust coefficient, and the total stiffness in the sway direction is determined by the contribution of the platform hydrostatic recovery and the contribution of the tendon system.

[0019] Furthermore, the average wave drift force in E4 is calculated using an empirical estimation relationship, which satisfies the following: the average wave drift force is proportional to the seawater density, gravitational acceleration, the square of the design wave amplitude, the empirical drift force coefficient, and the characteristic diameter of the platform, wherein the design wave amplitude is determined by the design wave height.

[0020] The present invention also provides a collaborative design optimization system for tension leg type floating wind turbine platforms for implementing the above method, comprising:

[0021] The design space definition and variable mapping module is used to establish automatic mapping of independent design variables and dependent variables and generate candidate designs;

[0022] Parametric modeling and mesh generation module, used to generate 3D geometric models and surface meshes;

[0023] The hydrostatic-mass-ballast matching module is used to calculate hydrostatic and mass characteristics and configure ballast in conjunction with the set tendon pretension.

[0024] The frequency domain hydrodynamic calculation module is used to calculate added mass, potential flow damping, hydrostatic recovery characteristics, and wave excitation.

[0025] The fast response evaluation module is used to evaluate the average drift displacement and solve for the natural period;

[0026] The composite penalty fitness genetic optimization module is used to construct fitness based on the objective function and constraints and drive population iteration.

[0027] The results output and post-processing module is used to output the globally optimal design and generate an evaluation report and a 3D model file.

[0028] Furthermore, the composite penalty fitness genetic optimization module uses real number encoding and employs crossover, mutation, and selection mechanisms for global optimization. Its fitness is constructed as the reciprocal of the sum of the estimated cost and the total penalty, where the total penalty is calculated separately by weighted summation of the violation amounts of periodic constraints, displacement constraints, and tendon-related constraints, and the penalty for each constraint is accumulated by the square of the relative violation degree exceeding the target or allowable range. The module also monitors the rate of change of the optimal fitness of the population, and when it falls below a preset threshold for several consecutive generations, it is determined to have converged and the iteration is terminated early.

[0029] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0030] Compared with existing technologies, the technical solution of this invention establishes a unified, automated, collaborative design closed loop, enabling changes in geometric parameters to synchronously drive the linked updates of hydrostatic, hydrodynamic, motion response, and cost indicators. This fundamentally eliminates the repeated backtracking problems caused by the fragmentation of various disciplines in traditional design processes. This mechanism allows key dynamic constraints such as periodic avoidance and average displacement to be quantified and incorporated into the optimization decision-making process early in the design phase, thereby significantly improving design efficiency and avoiding the instability of solutions caused by relying on experience-based adjustments. This achieves global optimization of the tension leg floating wind turbine platform in terms of engineering feasibility and economy.

[0031] Building upon this foundation, the technical features further enhance the following benefits: parametric modeling and automatic mesh generation improve the consistency between scheme generation and evaluation; the fast average drift evaluation mechanism significantly reduces computational costs while ensuring engineering accuracy; and the composite penalty fitness function and adaptive convergence strategy enhance the robustness and convergence stability of the optimization process under multiple constraints. These features, without altering the core technical principles, further improve the overall method's engineering applicability, computational efficiency, and industrial application value, making this invention more suitable for engineering applications and industrial implementation under complex working conditions.

[0032] (1) The expected benefits and commercial value of the technical solution of this invention after transformation are as follows:

[0033] This invention can significantly shorten the R&D cycle of tension leg floating wind turbine platforms in the conceptual and preliminary design stages, and effectively reduce design and verification costs. By integrating parametric modeling, multidisciplinary performance evaluation, and optimization solution processes into a unified automated closed loop, this invention can shorten the traditional design cycle, which relies on manual experience and repeated backtracking, from months to days or even hours, thereby significantly improving engineering design efficiency.

[0034] At the commercial application level, the technical solution of this invention can directly serve the early-stage scheme comparison and engineering decision-making phases of large-capacity floating wind power projects in deep-sea areas, providing wind power developers, turbine manufacturers, and engineering design units with a low-cost, high-reliability design support tool. By achieving synergistic optimization of the floating body and mooring system within a larger design space, this invention helps to obtain a tension leg floating wind turbine platform solution with lower total life-cycle costs and superior dynamic performance, thereby improving the economic feasibility and market competitiveness of floating wind power projects.

[0035] Furthermore, this invention employs open-source computing tools and an automated algorithm framework, possessing excellent engineering portability and industrial expansion potential. It can be further developed into specialized design software or design service platforms for floating wind power projects, exhibiting significant software product value and engineering service promotion value. It is expected to generate considerable economic benefits in floating wind power project design, scientific research consulting, and industrial applications.

[0036] (2) The technical solution of this invention fills a technical gap in the industry both domestically and internationally:

[0037] Existing research on tension leg floating wind turbine platforms, both domestically and internationally, largely focuses on single-disciplinary analysis or local optimization of a limited number of geometric parameters. It typically employs a phased, sequential design approach, starting with geometric design and then moving to dynamic calibration. Currently, there is no comprehensive technical solution that can unify and parametrically couple platform geometric parameters, hydrostatic characteristics, hydrodynamic response, mooring stiffness, and cost models during the conceptual design phase, and then use automated optimization algorithms to systematically optimize the entire design space.

[0038] This invention, for the first time in the field of tension leg floating wind turbine platform design, proposes and implements a fully automated collaborative design optimization method covering parametric modeling, automated multidisciplinary evaluation, constraint-driven optimization, and automatic result output. It incorporates mooring system parameters as endogenous design variables, integrating them along with the floating body's geometric parameters into a unified optimization framework, achieving truly integrated design of the floating body and mooring system. This technical approach overcomes the limitations of existing research that separates the design of the floating body and mooring system, filling a technological gap both domestically and internationally in fully automated collaborative optimization methods for tension leg floating wind turbine platforms.

[0039] (3) The technical solution of the present invention solves a technical problem that people have long wanted to solve but have never been able to solve successfully:

[0040] In the engineering practice of tension leg floating wind turbine platforms, how to simultaneously meet multiple constraints such as inherent period avoidance, average displacement control, tendon safety, and cost optimization during the conceptual design phase has always been a long-standing but difficult-to-solve technical challenge in the industry. Traditional design methods generally rely on engineering experience and repeated manual adjustments. When period conflicts or displacement exceedances occur during the subsequent verification phase, it is often necessary to significantly backtrack and reset the geometric parameters, resulting in low design efficiency and difficulty in ensuring global optimization.

[0041] This invention directly embeds inherent periodicity solving, rapid average drift assessment, and tendon constraints into an automated optimization process. This transforms periodic avoidance and displacement control from post-design verification conditions into core constraints that can be quantified and used in decision-making from the initial design stage. This fundamentally solves the long-standing technical dilemma of simultaneously achieving feasibility and economy in the design of tension leg floating wind turbine platforms. This technical solution makes it possible to quickly obtain engineering-feasible, dynamically safe, and cost-effective design solutions under complex constraints, successfully overcoming a key engineering problem that has long existed in the industry but is difficult to solve using traditional methods.

[0042] (4) The technical solution of the present invention overcomes technical bias:

[0043] There is a common technical bias in the existing technology, which is that the high-fidelity dynamic performance evaluation of tension leg floating wind turbine platforms must rely on time-domain coupled simulation methods with extremely high computational costs. As a result, it is difficult to carry out large-scale parameter optimization in the conceptual design stage, which often leads to the optimization work being postponed to the detailed design stage, or even replaced by experience-based judgment.

[0044] This invention constructs a rapid evaluation system based on a combination of frequency domain hydrodynamic analysis and quasi-static average drift assessment. While ensuring accuracy in critical engineering projects, it significantly reduces the computational cost of a single scheme evaluation, thus breaking the traditional misconception that high accuracy inevitably leads to low efficiency. Simultaneously, this invention introduces a composite penalty fitness function and an adaptive convergence strategy, enabling stable convergence of multi-constraint optimization problems within an automated framework, overcoming the industry's inherent bias that complex engineering problems are difficult to solve reliably through automated optimization.

[0045] Through the above-mentioned technical means, this invention proves that in the design of tension leg floating wind turbine platforms, efficient, reliable and engineering-guiding collaborative optimization can be achieved in the conceptual design stage, providing a new technical path for floating wind power engineering design methods that is different from the traditional experience-driven mode. Attached Figure Description

[0046] Figure 1This is a flowchart of the collaborative design and optimization method for tension leg floating wind turbine platform based on full-process automation provided in this embodiment of the invention;

[0047] Figure 2 This is a module diagram of the collaborative design and optimization system for tension leg floating wind turbine platform based on full-process automation provided in this embodiment of the invention;

[0048] Figure 3 This is a schematic diagram of a tension leg-type platform mesh model corresponding to a design scheme provided in the optimization process of this invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0050] In existing floating wind turbine platform engineering practices, the overall design of tension leg platforms has long relied on a phased analysis process dominated by manual experience, resulting in highly fragmented models across different disciplines. Geometric schemes are typically determined first, followed by separate hydrostatic, hydrodynamic, and structural checks. If, in later stages, cyclic avoidance, displacement, or tendon stress requirements are found to be unsatisfactory, manual reversal and repeated adjustments of geometric parameters are often necessary. This sequential design approach exhibits significant efficiency bottlenecks as platform dimensions increase and constraints multiply, with design cycles sometimes lasting months. Furthermore, it struggles to guarantee that the resulting solution achieves global optimization in terms of cost, dynamic response, and safety margin. More importantly, traditional methods fail to systematically explore the design space, easily falling into local empirical optima, and cannot support the current industrial demands for low-cost, high-reliability large-capacity wind turbines.

[0051] To address the aforementioned issues, this invention uses parametric design vectors as the core link, unifying geometric modeling, hydrostatic calculations, hydrodynamic analysis, motion response assessment, and cost estimation into a single automated evaluation chain. Changes in geometric parameters are first reflected in the platform's buoyancy and mass distribution in real time through the parametric modeling mechanism. At this stage, the hydrostatic module performs linked calculations of displacement, waterline characteristics, and center of buoyancy position, and adjusts the platform's overall center of gravity and tendon pretension state through an automatic ballast configuration mechanism. This process ensures that each candidate scheme is in a achievable static equilibrium state before entering dynamic analysis, avoiding the invalid calculations of numerous unusable schemes in traditional processes.

[0052] Based on this, the frequency domain hydrodynamic module performs radiation-diffraction solutions on the updated geometry and floating state to obtain the added mass, potential flow damping, and wave excitation loads. These hydrodynamic characteristics, along with the still-water restoring stiffness and tendon linearized stiffness, constitute the dynamic foundation of the platform-mooring coupled system. By solving for the eigenvalues ​​of undamped free vibration, the system can directly capture the response law of the platform's six-degree-of-freedom natural period as a function of geometry and stiffness, achieving direct coupling between periodic avoidance constraints and geometric design. This mechanism avoids traditional design methods that rely on empirical formulas or a posteriori verification, allowing periodic safety to be quantitatively controlled early in the design process.

[0053] Meanwhile, addressing the practical need for rapid assessment in engineering applications, the platform's average drift response under the combined effects of wind and waves is estimated using a quasi-static equivalent method. The average wind thrust and average wave drift force are introduced as a resultant force at the load level, directly related to the longitudinal total stiffness. This allows for an effective determination of the platform's steady-state offset level without introducing high-order time-domain calculations. This rapid assessment result, along with periodic constraints and tendon strength constraints, is incorporated into the optimization criteria, forming the decision-making basis for the synergistic effect of multiple constraints.

[0054] Ultimately, the genetic optimization mechanism uses the cost function as the objective and dynamic and structural safety constraints as the selection criteria to conduct an adaptive evolutionary search within the entire design space. The various modules do not operate in isolation, but rather form a closed-loop feedback loop through design variables and evaluation results. This allows geometry, load, response, and cost to mutually constrain and converge collaboratively in the same iterative process, thereby achieving system optimization of the tension leg floating wind turbine platform in terms of engineering feasibility and industrial economics.

[0055] like Figure 1 As shown in the figure, this invention provides a collaborative design optimization method for tension leg type floating wind turbine platforms based on full-process automation. The method specifically includes:

[0056] S1: Determine the optimization variables and design space: independent design variables, variable range, and dependent variables;

[0057] S2: Initialization: Randomly generate the initial population within the design space;

[0058] S3: Parametric geometric modeling and mesh generation: Generate the platform's 3D geometric model and surface mesh model;

[0059] S4: Automatic calculation of hydrostatic and mass characteristics: Calculates parameters such as drainage volume and matches ballast according to static balance;

[0060] S5: Frequency domain hydrodynamic coefficient calculation: Solve the radiation-diffraction problem to obtain matrices such as added mass, potential flow damping, and hydrostatic restoring force stiffness;

[0061] S6: Rapid assessment of average drift displacement: Obtain the average wind force by interpolation from the table, and obtain the average wave drift force by formulas for the pressure integral and viscosity term;

[0062] S7: Solving the system's equations of motion and natural periods: Solving the eigenvalues ​​of the undamped free vibration equation to obtain the natural periods of motion;

[0063] S8: Optimize the mathematical problem model: take the estimated cost as the objective function, and use periodic avoidance, displacement limit, tendon tension, cross-sectional strength, and minimum drainage as constraints.

[0064] S9: Composite penalty fitness function design: By adding a penalty term for not meeting the constraints, the problem of minimizing cost is transformed into the problem of maximizing fitness.

[0065] S10: Determine the genetic algorithm configuration and adaptive strategy: real number encoding, crossover and mutation, selection, and adaptive convergence framework;

[0066] S11: Evaluate the loop and evolutionary iteration. If the convergence condition is met, proceed to S12; otherwise, return to S3.

[0067] S12: Results Output and Post-processing: Output the globally optimal individual from each iteration, generate a detailed evaluation and design report, and output the 3D model file of the optimal design.

[0068] The collaborative design and optimization method for tension leg floating wind turbine platforms based on full-process automation described in this invention integrates parametric geometric modeling, hydrodynamic calculation, system dynamics analysis, and intelligent optimization algorithms to achieve collaborative optimization among platform structure, hydrodynamics, mooring system, and cost. Its overall working principle is as follows.

[0069] First, in step S1, the platform's geometric parameters, tendon arrangement parameters, and ballast parameters are determined as independent design variables, and their value ranges and dependencies are set, thereby constructing a complete design space. Step S2 involves randomly generating an initial population within this design space to enable parallel exploration of diverse design schemes.

[0070] Subsequently, in step S3, a three-dimensional parametric geometric model of the platform is automatically generated based on the design variables, and the surface mesh is generated simultaneously, providing a unified data interface for subsequent hydrodynamic analysis. In step S4, the hydrostatic and mass characteristics of the geometric model are calculated, the drainage volume, the position of the center of buoyancy, and the mass distribution are automatically obtained, and the platform achieves static equilibrium through matching ballast.

[0071] In step S5, the radiation-diffraction problem of the platform is solved based on potential flow theory to obtain hydrodynamic parameters such as the added mass matrix, potential flow damping matrix, and hydrostatic restoring force stiffness matrix. In step S6, table lookup interpolation is performed using a pre-established aerodynamic and hydrodynamic database, and the average drift force under wind and waves is quickly estimated by combining the pressure integral method and the viscous empirical model, thereby significantly reducing the computational cost of direct simulation from the frequency domain to the time domain.

[0072] In step S7, the obtained hydrodynamic parameters are substituted into the system motion equations to solve the eigenvalue problem of undamped free vibration, and the natural period of the platform in each degree of freedom is obtained to determine whether the design requirements for avoiding the dominant wave frequency are met.

[0073] In step S8, an optimization mathematical model is constructed with the estimated total cost of the platform as the objective function, and engineering constraints such as periodic avoidance, drift displacement limits, tendon tension limits, structural strength, and minimum drainage are introduced. In step S9, a composite penalty function is used to penalize individuals that violate the constraints according to the degree of constraint violation, thereby transforming the original constrained optimization problem into an unconstrained fitness maximization problem suitable for solving by a genetic algorithm.

[0074] In step S10, a real-valued genetic algorithm is configured, and an adaptive crossover rate and mutation rate adjustment mechanism is introduced to balance global search capability and local convergence speed. In step S11, the design solution set is continuously updated by iteratively evolving and evaluating the fitness of the population until the convergence criterion is met. Finally, in step S12, the globally optimal design scheme is output, and the corresponding 3D model file and comprehensive evaluation report are generated.

[0075] Through the above process, this invention achieves a fully automated closed loop from structural parameter definition to performance evaluation and optimization decision-making, improving design efficiency and ensuring engineering feasibility and economic optimality under multiple physical constraints.

[0076] S1 includes:

[0077] Independent design variables: Six geometric parameters with the highest sensitivity to the platform's hydrodynamic performance and manufacturing cost were selected as optimization variables, forming the design vector X. Typical variables include:

[0078] Draft (Unit: m)

[0079] Lower column height (Unit: m)

[0080] Lower column diameter (Unit: m)

[0081] upper column diameter (Unit: m)

[0082] Float length (Unit: m)

[0083] pontoon width (Unit: m)

[0084] Variable range: Set reasonable upper and lower limits for each independent variable. To form the initial design space .For example: .

[0085] Calculation of dependent variables: other geometric parameters, such as the height of the upper column. Location of tendon connection points Tendon length The results are automatically calculated by associating them with predefined geometric relationships and independent variables.

[0086] S2, in the defined design space Initial population is generated randomly within the population. ,scale .

[0087] The S3, based on the design vector X, uses a script written with the Python OCC library to automatically generate a three-dimensional geometric model of the platform, including components such as columns and pontoons.

[0088] The Gmsh library is used to automatically mesh the generated geometric model, producing a surface mesh file suitable for boundary element method calculations in potential flow theory. The mesh size can be adaptively set according to the wave frequency range to ensure calculation accuracy.

[0089] S4, based on the geometric model, automatically calculates the platform's drainage volume. Waterline surface area Floating Heart and drifting heart .

[0090] Based on the component geometry and the preset steel density, such as the upper column Lower column ,float Automatically calculate the weight of structural steel Through the static equilibrium equations Automatically distribute ballast weight To satisfy the given initial tendon pretension ,in This refers to the total pretension of the tendon.

[0091] S5 automatically submits the mesh file to the open-source potential flow solver Capytaine to solve the radiation-diffraction problem within the framework of linear potential flow theory.

[0092] Calculate the frequency Related additional mass matrix Potential flow damping matrix Wave excitation force / torque vector .

[0093] static water restoring stiffness matrix It is given by the following formula:

[0094]

[0095] in Let be the moment of inertia of the waterline.

[0096] The motion equation of the platform in the frequency domain (ignoring damping) in S7 is as follows:

[0097]

[0098] in, For the platform quality matrix, This is the linearized stiffness matrix of the tendon system.

[0099] By solving the eigenvalue problem of the undamped free vibration equation, the natural periods of the six degrees of freedom are obtained. This is the key constraint.

[0100] S6 refers to the average sway displacement of the platform under the combined action of wind and waves. This is the core constraint in the design of the tension leg platform. A simplified formula is used for rapid estimation:

[0101]

[0102] in Total stiffness in the sway direction (anchor stiffness) (Mainly).

[0103] Mean wind thrust According to rated wind speed Wind turbine thrust coefficient curve The mean wave drift force was obtained by interpolation from a table. Using wave height ,cycle Estimation of the design wave pressure integral and viscous term using empirical formulas:

[0104]

[0105] in, To design the amplitude, The empirical drift force coefficient, The diameter is the characteristic diameter of the platform.

[0106] The objective function of S8 is to minimize the estimated cost throughout the platform's entire lifecycle. .

[0107]

[0108] in This is the unit cost coefficient.

[0109] Constraints:

[0110] Periodic avoidance: , wait.

[0111] Displacement Limitation: .

[0112] Tendon tension: .

[0113] Cross-sectional strength: The cross-sectional area of ​​the tendon must meet the following requirements. (Tensile strength) and API RP 2T buckling resistance requirements.

[0114] Minimum displacement: .

[0115] S9, to efficiently handle the above constraints in the genetic algorithm, constructs the following fitness function. :

[0116]

[0117] in, The total penalty term is defined as follows:

[0118]

[0119] , These represent the sets of periodic constraints and tendon constraints, respectively.

[0120] , The first The periodic constraint and the first The amount of violation of a tendon constraint.

[0121] This is an adjustable penalty weighting coefficient used to balance the importance of different constraints.

[0122] Function Mechanism: This function transforms the original cost minimization problem into fitness maximization. The problem is that any violation of the constraints will be addressed through penalties. A sharp increase in the denominator significantly reduces the fitness of the individual, leading to its elimination through evolution. The squared term ensures a stronger penalty for serious violations of the constraints.

[0123] The S10 uses the openGA open-source library, and the design variables are encoded with real numbers.

[0124] Crossover and mutation employ simulated binary crossover (SBX) and polynomial mutation to balance global exploration and local exploitation.

[0125] The selection mechanism employs a tournament selection and elite retention strategy to ensure the inheritance of outstanding genes.

[0126] Adaptive convergence strategy: Monitor the rate of change of the optimal fitness of the population If continuous generation ,For example If convergence is achieved, the optimization is terminated early to save computational resources.

[0127] like Figure 2 As shown in the figure, the collaborative design and optimization system for tension leg floating wind turbine platforms based on full-process automation provided by this invention specifically includes:

[0128] The design space definition and parametric modeling module is used to transform engineering problems into optimization problems that can be handled by algorithms, and to establish an automatic mapping between geometric models and design variables;

[0129] An automated multidisciplinary performance evaluation module is used to build an integrated evaluation function. For any given design X, its hydrodynamic performance, motion response and cost can be automatically calculated;

[0130] The composite penalty fitness genetic algorithm optimization module is used to design a fitness function that can intelligently handle multiple objectives and constraints, so as to drive the genetic algorithm to find the best solution efficiently.

[0131] The fully automated optimization execution steps of the collaborative design optimization method for tension leg floating wind turbine platforms based on full-process automation provided in this embodiment of the invention are as follows:

[0132] (1) Initialization: in the defined design space Initial population is generated randomly within the population. ,scale .

[0133] (2) Evaluation cycle

[0134] For each individual in the population (design X), the complete evaluation process is automatically triggered, resulting in... and the amount of constraint violation. Based on the fitness function. Calculate the fitness of each individual.

[0135] (3) Evolutionary iteration

[0136] Perform selection, crossover, and mutation operations to generate offspring populations.

[0137] The offspring population is merged with the parent elites to form a new generation population.

[0138] Repeat the evaluation loop and evolutionary iteration until the convergence condition is met and the maximum number of generations is reached. Or it may trigger an adaptive convergence strategy.

[0139] (4) Results output and post-processing

[0140] Output the globally optimal individual in each iteration. .

[0141] Automatically generated includes A detailed design report including all geometric parameters, performance indicators (natural period, drift displacement, tendon tension), cost breakdown, and constraint satisfaction. It can automatically output a 3D model file or engineering drawings of the optimal design for subsequent detailed design or simulation verification.

[0142] In specific embodiments of this invention, the overall technical solution revolves around a closed-loop process of parametric modeling—multidisciplinary automatic evaluation—constraint-driven optimization—result output. Abstract functional expressions such as parametric modeling, automatic evaluation, and optimization solution are used. The above implementation methods are not limited to a single implementation path but are unfolded through multiple implementation levels: at the point level, specific executable tools, algorithms, or parameter value examples are provided; at the line level, the sequence of modules, data flow, and triggering relationships are clarified to form a stable and reproducible computational flow; at the surface level, alternative implementation methods and parameter range coverage demonstrate that the technical solution does not depend on a single software or a single numerical point. For example, parametric geometric modeling can be automatically generated by a scripted geometric kernel or implemented through a secondary development interface of a commercial modeling platform; mesh generation can employ a general surface element discretization method or a high-quality surface element partitioning strategy specifically for potential flow boundary elements.

[0143] This implementation discloses a multi-point coverage strategy of lower limit—median—upper limit to avoid the technical effect being misunderstood as being valid only at a single isolated numerical point. For example, at the optimization algorithm level, the initial population size can be preferably set to 50, 150, or 300 individuals to adapt to rapid exploration, medium-scale search, and high-dimensional complex design spaces, respectively; the crossover probability can be preferably set to 0.70, 0.85, or 0.95, and the mutation probability can be preferably set to 0.01, 0.05, or 0.20 to balance convergence speed and global search capability; the number of frequency discrete points in the frequency domain hydrodynamic calculation can be preferably set to 30, 60, or 120 to balance computational efficiency and response accuracy. Similarly, for constraint thresholds such as structural parameters, tendon pretension range, and allowable displacement ratio, multiple exemplary intervals or representative value points are given, and the applicable scenarios of these values ​​in engineering practice are clearly defined.

[0144] Regarding geometric model output, the model files can be exemplarily disclosed in various industry-standard formats such as STEP, IGES, or STL; at the mesh level, the element files can be output in a general data format or the native format of the potential flow solver; in hydrostatic and mass calculations, in addition to the drainage volume, the waterline moment of inertia, the mass matrix construction method, and the calculation paths for the center of buoyancy and the center of drift are preferably disclosed; regarding ballast configuration, various implementation methods such as concentrated ballast, compartmentalized ballast, or ballast continuously distributed along the structure are exemplarily provided; in load assessment, the mean wind thrust can be determined using linear interpolation, spline interpolation, or piecewise polynomial fitting, and the mean wave drift force can simultaneously consider the pressure integral term and the viscous correction term.

[0145] Example 1: Fully Automated Co-optimization Based on Scripted Geometric Modeling and Potential Flow Solver

[0146] In this embodiment, the independent design variables in the design space include draft, column height, column diameter, and pontoon dimensions. A three-dimensional geometric model of the tension leg floating wind turbine platform is automatically generated using scripted geometric modeling. The geometric modeling process is parameter-driven, and the geometric constraints between structural components are automatically established during the modeling phase. The generated three-dimensional model is then automatically discretized into a surface mesh and submitted to the frequency domain hydrodynamic solution module for calculating added mass, potential flow damping, hydrostatic recovery characteristics, and wave-induced loads. Based on the hydrodynamic and hydrostatic calculation results, the average drift displacement and natural period are further evaluated, and a genetic algorithm drives multi-generation iterative optimization to ultimately output the optimal platform design scheme that satisfies the constraints.

[0147] In this embodiment, the scripted geometric modeling tool can be replaced with different geometric kernels, and the method of generating surface meshes can be adjusted according to the required computational accuracy. The values ​​of relevant design variables cover a reasonable engineering range. This embodiment fully embodies the technical closed loop of parametric modeling—automatic evaluation—evolutionary optimization.

[0148] Example 2: Optimized Example Based on Multi-Interval Parameter Coverage

[0149] In this embodiment, a multi-interval coverage approach is used for optimization calculations on the geometric variables and algorithm parameters involved in the design space. Independent design variables are sampled with multiple representative values ​​between their upper and lower limits to form an initial population. During the optimization process, geometric modeling, hydrostatic calculations, hydrodynamic assessments, and constraint determination steps are performed on different individuals to ensure that the design scheme can be stably implemented within different parameter ranges. By comparing the cost and response indices under different parameter combinations, the genetic algorithm sequentially selects and retains individuals that meet the requirements of periodic avoidance and displacement constraints.

[0150] This embodiment clearly demonstrates that the method of the present invention does not rely on a single parameter point to achieve the expected technical effect, but is feasible and stable in a continuous range.

[0151] Example 3: Examples Based on Multiple Ballast Configurations

[0152] In this embodiment, the calculation of the platform's hydrostatic and mass characteristics includes not only the drainage volume and structural mass, but also the influence of different ballast configurations on the static equilibrium state. Ballast can be exemplarily arranged in a centralized manner inside the lower column, or it can be arranged in compartmentalized ballast or continuously distributed along the length of the pontoons. When calculating static equilibrium, the system automatically adjusts the ballast mass and distribution position according to the selected ballast method, so that the platform achieves a preset tendon pretension state while satisfying buoyancy balance.

[0153] This embodiment shows that the technical features of the ballast matching tendon pretension are not limited to a specific ballast structure or arrangement, but cover a variety of feasible engineering solutions.

[0154] Example 4: Example based on fast average drift assessment

[0155] In this embodiment, the average drift displacement of the platform under the combined action of wind and waves is obtained using a rapid evaluation method. The average wind thrust is determined by the rated wind speed and the turbine thrust coefficient through table lookup or interpolation, while the average wave drift force is estimated through empirical relationships. The system correlates the above average load with the total stiffness in the sway direction to obtain the average sway displacement of the platform, and uses it as one of the displacement constraints in the optimization judgment.

[0156] In this embodiment, the calculation method of average load can be simplified or refined according to engineering requirements, and the drift evaluation method does not affect the establishment of the overall optimization closed loop.

[0157] Example 5: An Example Based on Inherent Periodic Avoidance Constraints

[0158] In this embodiment, a coupled motion model of the platform and the tendon system is established to solve the eigenvalue problem of undamped free vibration and obtain the natural periods of six degrees of freedom. These natural periods are compared with the target wave period and the wind turbine's operating period to determine whether the period avoidance constraint is met. During the optimization iteration process, any design scheme that does not meet the period avoidance condition is penalized or directly eliminated, thereby guiding the optimization direction towards the dynamically safe region.

[0159] This embodiment clearly illustrates that solving for the inherent period and determining the constraints are important components of the method of this invention, but the specific mathematical solution method and the determination threshold can be adjusted within a reasonable range.

[0160] Example 6: Optimization Example Based on Composite Penalty Fitness Function

[0161] In this embodiment, the optimization objective is centered on the platform's estimated cost, while introducing multiple types of constraints. For constraints such as period, displacement, and tendon strength, a composite penalty term is constructed. When a design violates any constraint, the corresponding penalty term automatically increases, thereby reducing its fitness value. The genetic algorithm, driven by fitness maximization, prioritizes and performs crossover mutations on individuals that satisfy the constraints and have lower costs until the convergence condition is met.

[0162] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0163] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A collaborative design and optimization method for tension leg type floating wind turbine platforms based on full-process automation, characterized in that, The method includes: S1, determine the optimization variables and design space, establish the value range of independent design variables, and the automatic association between dependent variables and independent design variables; S2, generate an initial population within the design space; S3 performs parametric geometric modeling of individuals in the population and automatically generates surface meshes; S4 automatically calculates hydrostatic and mass characteristics, and determines the ballast load based on the static equilibrium relationship to match the given tendon pretension. S5, calculate the frequency domain hydrodynamic coefficients to obtain the added mass, potential flow damping, hydrostatic recovery characteristics and wave excitation; S6, based on the combined effect of wind and waves, to quickly assess the average drift displacement; S7. Based on the platform mass characteristics, added mass, hydrostatic recovery characteristics, wave excitation and tendon linearized stiffness, a motion model of platform-tendon coupling is established and the natural period is solved. S8. Construct an optimized mathematical model with the goal of cost estimation and constraints such as periodic avoidance, displacement limitation, tendon tension, cross-sectional strength and minimum drainage. S9. Construct a composite penalty fitness function to transform the cost minimization problem into a fitness maximization problem; S10, execute the genetic algorithm evolution iteration: repeat S3 to S9 for each generation of individuals and update the population, and terminate when the convergence condition is met; S11 outputs the globally optimal individual for each iteration and generates the corresponding evaluation report and the 3D model file of the optimal design; S5 includes: solving the radiation and diffraction problem within the framework of linear potential flow theory to obtain the additional mass, potential flow damping, and wave excitation that vary with frequency; and calculating the still water recovery characteristics, wherein the still water recovery characteristics provide recovery terms only for heave, roll, and pitch in the six degrees of freedom, and the heave recovery term is proportional to the waterline area, the roll recovery term is proportional to the moment of inertia of the waterline about the longitudinal axis, the pitch recovery term is proportional to the moment of inertia of the waterline about the transverse axis, and the still water recovery terms for the other degrees of freedom are zero.

2. The method as described in claim 1, characterized in that, S1 includes: the independent design variables include draft, lower column height, lower column diameter, upper column diameter, buoy length and buoy width; and upper and lower limits are set for each independent design variable to form an initial design space, and the remaining geometric parameters are automatically calculated by associating the predefined geometric relationships with the independent design variables.

3. The method as described in claim 1, characterized in that, S3 includes: automatically generating a three-dimensional geometric model of the platform based on the independent design variables, the three-dimensional geometric model including columns and pontoon components; and automatically performing surface meshing on the three-dimensional geometric model to generate a surface mesh file suitable for boundary element hydrodynamic calculations of linear potential flow theory.

4. The method as described in claim 1, characterized in that, The S4 includes: automatically calculating the platform's drainage volume, waterline area, buoyancy center position, and drift center position based on a three-dimensional geometric model; automatically calculating the structural mass based on the component's geometric dimensions and preset material density; and then determining the ballast weight in conjunction with the static equilibrium conditions to ensure that the buoyancy force on the platform is balanced with the platform's total weight and the total pretension of the tendons, and to ensure that the initial tendon pretension reaches the set value.

5. An automated performance evaluation method for collaborative design optimization of tension leg type floating wind turbine platforms, characterized in that, The system automatically outputs evaluation results for constraint determination and cost calculation for any given design scheme, wherein the evaluation method includes: E1 automatically generates the platform geometry model and generates surface mesh; E2 automatically calculates hydrostatic and mass characteristics and determines the ballast in conjunction with the static balance and satisfies the set tendon pretension. E3 automatically calculates frequency domain hydrodynamic coefficients to obtain added mass, potential flow damping, hydrostatic recovery characteristics and wave excitation. E4, quickly assess the average drift displacement; E5. Based on the platform mass characteristics, added mass, hydrostatic recovery characteristics, wave excitation and tendon linearized stiffness, a motion model of platform-tendon coupling is established and the natural period is solved. E6 outputs an evaluation result vector containing average drift displacement, natural period, tendon tension and strength criterion, displacement and cost, which can be used by the optimization module. In E4, the radiation and diffraction problems are solved within the framework of linear potential flow theory to obtain the additional mass, potential flow damping, and wave excitation that vary with frequency; and the still water recovery characteristics are calculated, wherein the still water recovery characteristics provide recovery terms only for heave, roll, and pitch in the six degrees of freedom, and the heave recovery term is proportional to the waterline area, the roll recovery term is proportional to the moment of inertia of the waterline about the longitudinal axis, the pitch recovery term is proportional to the moment of inertia of the waterline about the transverse axis, and the still water recovery terms for the other degrees of freedom are zero.

6. The method as described in claim 5, characterized in that, In E4, the average drift displacement is estimated by dividing the sum of the average wind thrust and the average wave drift force by the total stiffness in the sway direction. The average wind thrust is obtained by interpolation from the curve of rated wind speed and wind turbine thrust coefficient. The total stiffness in the sway direction is determined by the contribution of the platform hydrostatic recovery and the contribution of the tendon system.

7. The method as described in claim 6, characterized in that, The average wave drift force in E4 is calculated using an empirical estimation relationship, which satisfies the following: the average wave drift force is proportional to the seawater density, gravitational acceleration, the square of the design wave amplitude, the empirical drift force coefficient, and the characteristic diameter of the platform, wherein the design wave amplitude is determined by the design wave height.

8. A collaborative design optimization system for a tension leg type floating wind turbine platform for implementing the method as described in any one of claims 1 to 4, characterized in that, include: The design space definition and variable mapping module is used to establish automatic mapping of independent design variables and dependent variables and generate candidate designs; Parametric modeling and mesh generation module, used to generate 3D geometric models and surface meshes; The hydrostatic-mass-ballast matching module is used to calculate hydrostatic and mass characteristics and configure ballast in conjunction with the set tendon pretension. The frequency domain hydrodynamic calculation module is used to calculate added mass, potential flow damping, hydrostatic recovery characteristics, and wave excitation. The fast response evaluation module is used to evaluate the average drift displacement and solve for the natural period; The composite penalty fitness genetic optimization module is used to construct fitness based on the objective function and constraints and drive population iteration. The results output and post-processing module is used to output the globally optimal design and generate an evaluation report and a 3D model file.

9. The system as described in claim 8, characterized in that, The composite penalty fitness genetic optimization module uses real number encoding and employs crossover, mutation, and selection mechanisms for global optimization. Its fitness is constructed as the reciprocal of the sum of the estimated cost and the total penalty. The total penalty is calculated separately by weighting and summing the violations of periodic constraints, displacement constraints, and tendon-related constraints. The penalty for each constraint is accumulated by the square of the relative violation degree that exceeds the target or allowable range. The module also monitors the rate of change of the optimal fitness of the population. When the fitness is below a preset threshold for several consecutive generations, convergence is determined and the iteration is terminated early.

Images