A novel biomimetic scour protection system for monopile wind turbines and its optimization method

Abstract

This invention provides a novel biomimetic scour protection system and optimization method for monopile wind turbines, belonging to the field of offshore wind power foundation scour protection technology. The system includes: a protective skirt installed around the monopile of the monopile wind turbine foundation, using the Fibonacci spiral of a sunflower seed as a biomimetic prototype; and protective pins arranged in an alternating Fibonacci spiral pattern on the protective skirt to construct the initial configuration. This maximizes interaction, ensures continuous water flow deflection, minimizes wake effects, and maximizes turbulence and energy dissipation across the entire skirt.

Description

Technical Field

[0001] This invention relates to the field of offshore wind power foundation scour protection technology, and in particular to a novel biomimetic scour protection system and optimization method for monopile wind turbines. Background Technology

[0002] In recent years, the use and application of offshore wind energy has increased significantly. According to the Global Offshore Wind Energy Report, the offshore wind power industry added 8 GW of installed capacity in 2024, bringing the global total installed capacity to 83 GW. Recent advancements in turbine technology have resulted in larger, taller wind turbines with longer blades, significantly improving their power generation capacity. Among various offshore wind turbine foundations, monopile foundations are widely used in near-shore shallow-water offshore wind farms due to their simple structure, ease of construction, and cost-effectiveness. Statistics show that monopile foundations account for over 70% of operational offshore wind farms globally. The long-term stability of offshore wind turbines faces many challenges, among which scour around monopile foundations is crucial.

[0003] Scouring occurs because the monopile foundation disturbs the surrounding flow field, leading to turbulence, eddy shearing, and increased bed shear stress, which in turn causes sediment transport and localized scouring around the pile. The presence of a cylindrical monopile in the flow field generates horseshoe-shaped eddies, which eventually erode the soil and form localized scour pits near the structure.

[0004] Previous studies have focused on detecting scour around monopiles, including experimental studies on scour around monopiles and experimental studies on localized scour around square-section porous monopiles under constant flow velocity. Numerical simulations have also been conducted on the response to concentrated wave loads around monopil foundations. Another important scour protection measure is the implementation of nearshore protection measures. Scour protection systems are divided into active and passive methods. Passive protection focuses on altering the erosiveness of the bed surface or soil properties, primarily by laying erosion-resistant materials to improve its resistance. Some common passive protection measures include riprap revetments, gabions, and geotextile mats. These methods offer limited mitigation and often fail to function reliably under varying hydrodynamic conditions and long-term topographic evolution. On the other hand, active protection controls the water flow around the foundation, altering local hydrodynamic conditions to reduce erosion and thus suppressing scour. These methods include sheathing, sacrificial piles, trench piles, underwater dikes, and sand-blocking dikes.

[0005] Erosion protection has been studied using empirical, experimental, and numerical techniques. However, understanding the flow field and capturing the detailed flow distribution beneath the structure is extremely difficult, leaving erosion mitigation measures unclear.

[0006] Therefore, this invention proposes a novel biomimetic scour protection system for monopile wind turbines and an optimization method thereof. Summary of the Invention

[0007] This invention provides a novel biomimetic scour protection system and optimization method for monopile wind turbines to solve the aforementioned technical problems.

[0008] This invention provides a novel biomimetic scour protection system for monopile wind turbines, comprising: using the Fibonacci spiral of sunflower seeds as a biomimetic prototype, setting a protective skirt around the monopile of the monopile wind turbine foundation, and arranging protective pins in an alternating Fibonacci spiral pattern on the protective skirt to construct an initial configuration, which is regarded as a novel biomimetic scour protection system for monopile wind turbines.

[0009] This invention provides an optimization method for a novel biomimetic scour protection system for monopile wind turbines. Based on the aforementioned novel biomimetic scour protection system for monopile wind turbines, the following steps are performed, specifically including: Step 1: Determine the core design variables and set the value range for each design variable; Step 2: Experimental design was carried out using an orthogonal array to construct a three-dimensional CFD simulation model that couples the immersion boundary method with the bed sand morphology and dynamics model, and to perform single-pile scour protection simulation calculations on the combination of design variables. Step 3: Based on the simulation calculation results, conduct correlation matrix analysis, regression analysis and response optimization to determine the optimal combination of design variables for the novel biomimetic scour protection system for monopile wind turbines based on the pre-set optimization objectives; Step 4: The protective skirt corresponding to the optimal design variable combination is divided into three interlocked modules. The modules are spliced ​​using a Gaussian sine key radial interlock structure to obtain the final optimized monopile wind turbine biomimetic scour protection system.

[0010] Preferably, the mesh independence is verified by the mesh convergence index GCI, and a medium mesh size is selected for simulation calculation; The grid convergence index (GCI) is calculated as follows: ; ; Where p is the order of convergence; r is the constant refinement ratio; It is the convergence result of the scouring depth of the fine mesh; This is the convergence result of the scour depth in the medium grid; This is the convergence result of the scour depth of the coarse mesh; This is relative error; is the confidence factor.

[0011] Preferably, the protective skirt corresponding to the optimal design variable combination is divided into three interlocking modules, and the three interlocking modules are three 120° equal parts.

[0012] The preferred and final optimized monopile wind turbine biomimetic scour protection system is made of high-performance fiber-reinforced concrete. The modular components are prepared through a two-stage volume casting process, 50Hz vibration compaction for 120 seconds, 24-hour ambient temperature curing, and 48-hour immersion hydration.

[0013] Preferably, before assembling the modules based on the optimal combination of design variables, the method further includes: The protective skirt corresponding to the optimal design variable combination is decomposed, and a global quantitative formula for the multi-physics integrated anomaly index of the decomposed blocks is established. : ,in, The multiphysics comprehensive anomaly index for the i1th split block; The relative deviation of the shear stress on the bed surface surrounding the i1th split block; The critical starting shear stress of the bed sand; The stress deviation of the Gaussian sinusoidal interlocking surface structure of the i-th split block; To achieve the allowable stress of high-performance fiber-reinforced concrete; The geometric profile tolerance deviation of the i1th split block; Tolerances are allowed for modular geometry; The deviation of the hydrodynamic continuity distortion coefficient for the i1th split block; To design the hydrodynamic continuity factor; , , , The weighting coefficients of the analytic hierarchy process (AHP)-entropy weight composite system satisfy the following conditions: ; Traverse the three split blocks If all split blocks If the preset splitting scheme is deemed acceptable, then... The default threshold for abnormal indicators is set. Otherwise, extract the multiphysics simulation anomaly architecture for each split block.

[0014] Preferably, after extracting the simulation anomaly architecture of each split block, it also includes: Construct a set of topological anomaly feature vectors and determine the anomaly degree of the improved information entropy-grey relational degree composite architecture: ,in, denoted as the composite architecture anomaly degree of the i1th split block; m is the total number of anomaly types in the topological anomaly feature vector set; Let be the normalized proportion of the j-th type of outlier in the i1-th split block, and ; The value of the j-th type of anomaly index for the i1th split block; is the gray resolution coefficient, with a value of 0.1 to 0.5; k is the segmentation block number, with a value of 1, 2, or 3; Anomaly based on composite architecture Construct an augmented mapping matrix for abnormal parameters, and determine the diagonal matrix of the split block structure and hydrodynamic added parameters. : ,in, For parameter correction coefficients; These are the radial span, circumferential arc length, and axial height of the protective skirt, respectively. With the minimization of composite anomaly degree, optimal matching of added parameters, maximization of interlocking structural stiffness, and optimal continuity of hydrodynamic field as the core objectives, a new objective function with multiple constraints is constructed. : ,in, , , It is a multi-objective balance coefficient, and it belongs to the range of 0 to 1; Let be the stiffness of the interlocking structure of the i1th split block, and ; Let be the hydrodynamic continuity coefficient of the i1th split block, and ; ,and This represents the maximum permissible value for the anomaly degree of the composite architecture. This is the baseline value for the radial span of the protective skirt; The design reference stiffness for the interlocking structure; Minimum allowable stiffness; This represents the maximum permissible value for the anomaly degree of the composite architecture.

[0015] Preferably, after constructing the multi-constraint objective function, it also includes: Based on the newly added objective function with multiple constraints, perform co-optimization of architecture and parameters for each split block: ,in, For the i1th split block in the tth iteration, the structure-hydrodynamic composite parameter vector is used. For adaptive iteration step size; To add the first-order partial derivative of the objective function; The Hessian matrix is ​​added as a new objective function; These are the regularization coefficients of the Hessian matrix; It is the identity matrix; Let be the Lagrange multiplier for the c-th constraint; Let c be the c-th constraint function; n is the total number of constraints. For the i1th split block in the (t+1)th iteration, the structure-hydrodynamic composite parameter vector is used. Iterate to When all constraints are met, a qualified three-stage interlock module is obtained, wherein, The iteration threshold; Based on the optimal parameter combination of the simulation qualified module, the automatic and precise disassembly of the protective skirt is achieved through a two-way mapping of geometric parameters and disassembly topology. Compared with the prior art, the beneficial effects of this application are as follows: Extremely high accuracy was achieved, verifying that the model's prediction error for the maximum scour depth was only 4.4%, consistent with benchmark data, thus establishing a reliable platform for systematic design exploration. Through comprehensive Design of Experiments (DOE) and subsequent multi-objective statistical optimization, key and non-intuitive design principles were extracted from 36 simulation configurations. The analysis clearly indicated the pin shape (…). ) is the most influential design factor, ( , This demonstrates that the system's effectiveness relies on its biomimetic flow customization, rather than simple batch addition. This optimization successfully yielded a robust and practical design that achieves 38.8% scour protection efficiency (SPE) while maintaining critical hydrodynamic neutrality. ).

[0016] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating an optimization method for a novel biomimetic scour protection system for a single-pile wind turbine, as described in an embodiment of the present invention. Figure 2 This is a framework diagram of the descriptive research model in an embodiment of the present invention; Figure 3 The diagrams are related to Model I of this embodiment of the invention; Figure 4 The relevant diagrams are for Model II of this embodiment of the invention; Figure 5 The relevant diagrams are for Model III of this embodiment of the invention; Figure 6 The diagram shows the relevant figures for Model IV of this embodiment. Detailed Implementation

[0019] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0020] This invention provides a novel biomimetic scour protection system for monopile wind turbines, comprising: using the Fibonacci spiral of sunflower seeds as a biomimetic prototype, setting a protective skirt around the monopile of the monopile wind turbine foundation, and arranging protective pins in an alternating Fibonacci spiral pattern on the protective skirt to construct an initial configuration, which is regarded as a novel biomimetic scour protection system for monopile wind turbines.

[0021] The dense, efficient arrangement of the sunflower seed array is transformed into a dense array of vertical pins (additional scour protection) mounted on the protective skirt, the positions of which strictly follow the Fibonacci spiral formula. This layout is crucial because adjacent rows of pins are always staggered. This staggered arrangement maximizes interaction with the incoming flow, ensures continuous flow deflection, minimizes wake effects, and maximizes turbulence and energy dissipation throughout the skirt.

[0022] This invention provides an optimization method for a novel biomimetic scour protection system for monopile wind turbines, such as... Figure 1 As shown, it includes: Step 1: Determine the core design variables and set the value range for each design variable, as shown in Table 1.

[0023] Table 1 Design Variables and Corresponding Ranges Step 2: Experimental design was carried out using an orthogonal array to construct a three-dimensional CFD simulation model that couples the immersion boundary method with the bed sand morphology and dynamics model, and to perform single-pile scour protection simulation calculations on the combination of design variables. Step 3: Based on the simulation calculation results, conduct correlation matrix analysis, regression analysis and response optimization to determine the optimal combination of design variables for the novel biomimetic scour protection system for monopile wind turbines based on the pre-set optimization objectives; In this embodiment, the scour protection system aims to meet multiple objectives, including maximizing scour protection efficiency, ensuring a bed shear stress ratio of no more than 1, maximizing the protection efficiency index, and minimizing the pile drag ratio, while ensuring high hydrodynamic performance and maintaining structural stability and cost-effectiveness. Scour protection efficiency = (1 - maximum scour depth with protection / maximum scour depth without protection) × 100%; Surface shear stress ratio = maximum bed shear stress / critical bed shear stress; Protection efficiency index = scour depth reduction / material cost, material cost is proportional to stone volume × number of stones; Pile drag force ratio = drag force of protected piles / drag force of unprotected piles.

[0024] Step 4: The protective skirt corresponding to the optimal design variable combination is divided into three interlocked modules. The modules are spliced ​​using a Gaussian sine key radial interlock structure to obtain the final optimized monopile wind turbine biomimetic scour protection system.

[0025] Preferably, the three-dimensional CFD simulation model is solved by sequential coupling of hydrodynamics and morphological dynamics, and the mesh independence is verified by the mesh convergence index GCI.

[0026] Preferably, the grid convergence index (GCI) is calculated as follows: ; Where p is the order of convergence; r is a constant refinement ratio; It is the convergence result of the scouring depth of the fine mesh; This is the convergence result of the scour depth in the medium grid; This is the convergence result of the scour depth of the coarse mesh; This is relative error; is the confidence factor.

[0027] The preferred and final optimized monopile wind turbine biomimetic scour protection system is made of high-performance fiber-reinforced concrete. The modular components are prepared through a two-stage volume casting process, 50Hz vibration compaction for 120 seconds, 24-hour ambient temperature curing, and 48-hour immersion hydration.

[0028] In this embodiment, a descriptive research model is used for discussion, such as... Figure 2 As shown, four hypothetical models were tested using correlation and regression analysis (using Minitab software). This analysis quantified the five predictor variables and four response variables (SPE, ...). , The relationship between (and PDFR). For minimizing objectives (e.g., ... (and PDFR) determines the optimal design level based on the principle that the smaller the mean, the better; for maximizing the objective (SPE and PDFR), the optimal design level is determined; The optimal design level is determined based on the principle that the larger the mean, the better.

[0029] Assumption Model I: (Variable: x) has a significant positive correlation with scour protection efficiency (SPE).

[0030] Assumption Model II: (Variable: x) relative to the bed shear stress ratio There is a significant negative correlation.

[0031] Assumption Model III: (Variable: x) has a significant positive correlation with the protection efficiency index (η).

[0032] Hypothesis Model IV: (Variable: x) has a significant negative correlation with the pile drag ratio (PDFR).

[0033] In this embodiment, E{SPE}, E{ }, E{η} and E{PDFR} are the response variable values ​​of models I, II, III and IV in the ith operation, respectively.

[0034] , , , which are the parameters of models I, II, III and IV, respectively.

[0035] Given variable x= is a constant, i.e., the value of the predicted variable in the i-th operation of models I, II, III, and IV.

[0036] random error term The mean is E{Y }=0, and the variances are respectively , .

[0037] In this embodiment, Pearson correlation analysis was performed to determine the linear relationship between five independent design parameters and four measured performance responses of the scour protection system. This analysis determined the strength (ρ) and statistical significance (p-value) of the association between each pair of variables. The correlation matrix provides an overview of the linear relationship between the S / N ratios of the response variables, with a value of 1.0 indicating a perfect positive correlation, -1.0 indicating a perfect negative correlation, and 0 indicating no linear correlation, as shown in Table 2. Table 2 Correlation among response variables As shown in Table 2 above, there is a highly significant positive correlation between scour protection efficiency (SPE) and protection efficiency index (η) (ρ=0.70).

[0038] The correlation analysis of scour protection efficiency (SPE) is shown in Table 3: Table 3 Correlation analysis of scour protection efficiency (SPE) Impact and discussion on scour protection efficiency (SPE): The pin shape (V3N) parameter is significantly positively correlated with SPE (ρ=0.732, p<0.001), while the effects of parameters V1, V2, V4, and V5 on SPE are negligible.

[0039] The correlation analysis of the bed surface shear stress ratio (τ*) is shown in Table 4: Table 4 Correlation Table of Bed Surface Shear Stress Ratio In this embodiment, the correlation between some parameters and τ* is weak or negligible. Among them, the negative correlation between the pin shape and τ* is the strongest (although still weak), which suggests that, to some extent, a better pin shape may slightly reduce the bed shear stress.

[0040] The correlation analysis of the protection efficiency index (η) is shown in Table 5: Table 5 Correlation Analysis of Protection Efficiency Index (η) The pin shape (V3N) was the only factor that had a statistical significance for η, and the two showed a moderate positive correlation (ρ=0.427, p=0.009), which confirms that the benefits of the pin shape are not only reflected in physical protection, but also in the overall efficiency of the system.

[0041] The pile drag ratio PDFR is used to measure structural loads and should ideally be minimized, as shown in Table 6. Table 6 Correlation Analysis of Pile Drag Ratio (PDFR) As shown in Table 7, at the α=0.05 level, no parameter showed significant correlation; however, the skirt diameter (V1) showed a weak positive correlation (ρ=0.286), which was marginally significant at the 10% level (p=0.093). This indicates that the pile drag ratio tends to increase slightly with increasing skirt diameter.

[0042] Table 7 summarizes the strongest and weakest correlations between design parameters and performance response. In this embodiment, respectively in Figure 3 , Figure 4 , Figure 5 and Figure 6 The average performance impact of each factor—scour protection efficiency (SPE), bed shear stress ratio (τ*), protection efficiency index (η), and pile drag ratio—is shown in the figure.

[0043] Model I: E(SPE) (scour protection efficiency): The overall regression model of E(SPE) is statistically significant (F value = 3.32, P value = 0.007).

[0044] Significant impact: Only the bidirectional interaction term V2*V4 showed a statistically significant impact on the scour protection efficiency (SPE) (P=0.036<0.05).

[0045] Main effects: No individual significance was found for any of the individual predictor variables (V1, V2, V3N, V4, V5). >0.05).

[0046] Visualization Analysis: As shown in the main effects plot, V3N exhibits the strongest positive effect on SPE, leading to a sharp increase in average SPE within its range. V5 shows a moderate negative trend, with its increase often resulting in a decrease in SPE. The interaction plot confirms the significant interaction V2*V4, with the lines in the plot showing a clear crossover pattern, indicating that the effect of V4 on SPE largely depends on the level of V2.

[0047] Model II: E( (Ratio of bed surface shear stress): E( The overall regression model was not statistically significant at the α=0.05 level (F=1.85, P=0.100). However, specific bidirectional interaction terms were found to have a significant effect on the bed shear stress ratio.

[0048] Significant effects: Two significant two-way interaction terms were found: ( -Value ). ( -Value ).

[0049] Marginal significance main effect: V1 has marginal significance (P=0.062), indicating that it has a relatively strong individual effect.

[0050] Visualization analysis: The main effects plot shows that V1 has the greatest impact, exhibiting a steep positive slope on the average bed shear stress ratio, while V5 shows a negative slope. The interaction plot visually confirms the significant interactions between V2*V4 and V2*V5, and these interactions present clear non-parallel lines, indicating the existence of conditional effects.

[0051] Model III: E(η) (protection efficiency index): The overall regression model of E(η) is not statistically significant (F value = 1.03, P value = 0.470).

[0052] Significant impact: At the significance level of α=0.05, no single predictor variable or two-way interaction term had a statistically significant impact on the protection efficiency index (all P values ​​> 0.05).

[0053] Visual analysis: Although the analysis of variance (ANOVA) lacked statistical significance, the main effects plot showed that V1 and V3N had the most significant effects. Increasing V1 appeared to decrease E( Increasing V3N seems to increase E( All interaction plots show that the curves for all factor combinations are almost parallel, which is consistent with the result that the interaction terms are not significant.

[0054] In this embodiment, the overall regression model of model IV: E(PDFR) (pile drag ratio): E(PDFR) is not statistically significant (F value = 1.42, P value = 0.228).

[0055] Significant effect: At the α=0.05 level, no single predictor variable or two-way interaction term had a statistically significant effect on the pile drag ratio (all P values ​​> 0.05).

[0056] The main effects with marginal significance were V1 (P=0.084) and V3N (P=0.083), indicating that they were close to the significance threshold. V2*V3N was also a marginally significant interaction term (P=0.074).

[0057] Visual analysis: The main effects plot clearly shows that V1 has a strong positive effect on PDFR, while V3N and V4 both have strong negative effects on PDFR. These main effects have the strongest visual impact among the non-significant models. The interaction plot shows several non-parallel lines (e.g., V1×V2 and V1×V5), which visually indicate the presence of interactions, even though they are not statistically significant at α=0.05.

[0058] In summary, Model I (SPE): The overall model is statistically significant (P=0.007), and only the V2*V4 interaction has a statistically significant effect (P=0.036). V3N shows the strongest main effect.

[0059] Model II The overall model was not significant (P=0.100); however, the two interactions, V2*V4 (P=0.034) and V2*V5 (P=0.017), were statistically significant. V1 had the strongest main effect (marginally significant, P=0.062).

[0060] Model III The overall model is not significant. ), and no single factor or two-way interaction was statistically significant ( V1 and V3N have the strongest main effects (negative and positive, respectively).

[0061] Model IV (PDFR): The overall model was not significant (P=0.228), and no single factor or two-way interaction was statistically significant (P>0.05). V1 and V3N had the strongest main effects (marginally significant, P=0.084 and P=0.083, respectively).

[0062] In this embodiment, the function for optimizing the objective is: j2 = 1, 2, 3, 4; Through a comprehensive analysis combining statistical regression and multi-objective response optimization, we gain a deeper understanding of the independent variables (V1 to V5) and four performance indicators (SPE, , The relationship between (and PDFR) was analyzed, and regression analysis showed that only the model of scour protection efficiency (SPE) was statistically significant. This indicates that the selected factors and their interactions can reliably predict changes in SPE.

[0063] Among all models, the effects of the two-way interaction terms were most significant, particularly V2*V4 and V2*V5. This indicates that the influence of one factor on the response variable largely depends on the levels of other interaction factors. The individual (main) effects of V1 and V3N showed the strongest influence in all four responses, typically appearing as marginally significant terms. The optimal setting for each response variable achieved a perfect composite expected value D=1.000, maximizing SPE and [missing value] under different specific factor settings. and minimize Along with PDFR, multi-response optimization yields an efficient compromise with a composite expectation of D=0.8638, balancing the conflicting objectives of maximization and minimization. To achieve balanced performance, the recommended optimal factor settings are: V1=3, V2=0.2, V3N=6 (streamlined pin shape), V4=0.2, and V5=0.551515.

[0064] In this embodiment, the splicing surface of the three-segment 120° equally divided module adopts a radially symmetrical Gaussian sine key interlocking structure. This structure achieves a gapless and precise fit between the concave and convex keys through a sine curve modulated by a Gaussian function, and also has the advantages of radial self-locking, resistance to wave cyclic loads, and waterproof and corrosion-resistant properties. In this embodiment, the specific parameters of the core geometric parameters of the interlocking structure are shown in Table 8: Table 8 Core geometric parameters of the interlocking structure In this embodiment, the Fibonacci spiral arrangement of the airfoil pin adopts a sunflower seed double helix array design, including two sets of Fibonacci spirals, one left-handed and one right-handed. The number of spirals is a combination of adjacent numbers in the Fibonacci sequence. The center positions of the airfoil pins are arranged in polar coordinates, and the core formula is: In the formula: The radial distance from the center of the airfoil pin to the center of the single pile, in mm; Let be the helical scaling factor, and take . =4.2mm, the engineering prototype is scaled down proportionally to the length scale; The serial numbers are for the airfoil pins, starting from 1 and increasing sequentially; The polar angle of the airfoil pin, in rad; The golden ratio is used to ensure the uniformity and interlacing of the spiral arrangement.

[0065] In this embodiment, the number of Fibonacci spirals adopts a combination of adjacent Fibonacci numbers of "8 left-handed spirals and 13 right-handed spirals". This combination is the natural optimal arrangement of sunflower seeds, achieving uniform staggering of the airfoil pins in the entire circumference without any blind spots. The impact of different spiral numbers on the protection efficiency is verified through pre-simulation, as shown in Table 9. Table 9. Effect of different spiral numbers on protection efficiency According to the verification results, the combination of 8 / 13 spirals is the optimal value. This parameter is incorporated into the initial biomimetic configuration design of this embodiment. The final total number of airfoil pins is 104, which completely covers the effective protection area of ​​the protective skirt. Adjacent airfoil pins are always staggered to maximize the dissipation of water flow energy and avoid local shear stress increase.

[0066] This invention provides an optimization method for a novel biomimetic scour protection system for monopile wind turbines. Before splitting the protective skirt corresponding to the optimal design variable combination into three interlocking modules, the method further includes: Establish a global quantitative formula for multi-physics integrated anomaly index of split blocks. : ,in, The multiphysics comprehensive anomaly index for the i1th split block; The relative deviation of the shear stress on the bed surface surrounding the i1th split block; The critical starting shear stress of the bed sand; The stress deviation of the Gaussian sinusoidal interlocking surface structure of the i-th split block; To achieve the allowable stress of high-performance fiber-reinforced concrete; The geometric profile tolerance deviation of the i1th split block; Tolerances are allowed for modular geometry; The deviation of the hydrodynamic continuity distortion coefficient for the i1th split block; To design the hydrodynamic continuity factor; , , , The weighting coefficients of the analytic hierarchy process (AHP)-entropy weight composite system satisfy the following conditions: ; In this embodiment, The critical threshold for bed sand initiation is determined by calculating the particle size, density, and Shields critical number of sediment in the project area. Based on the mechanical properties of materials, the allowable tensile stress is ≥4MPa and the allowable compressive stress is ≥70MPa under the specified mix proportions. Based on the recommended engineering value of ±0.5mm, it can be adjusted according to the accuracy requirements of offshore assembly; This is the baseline value under the optimal parameter combination, used to ensure that the water flow field is undistorted after splitting and to maintain hydrodynamic neutrality.

[0067] Traverse the three split blocks If all split blocks If the preset splitting scheme is deemed acceptable, then... The default threshold for abnormal indicators is set. Otherwise, extract the multiphysics simulation anomaly architecture for each split block.

[0068] Preferably, after extracting the simulation anomaly architecture of each split block, it also includes: Construct a set of topological anomaly feature vectors and determine the anomaly degree of the improved information entropy-grey relational degree composite architecture: ,in, denoted as the composite architecture anomaly degree of the i1th split block; m is the total number of anomaly types in the topological anomaly feature vector set; Let be the normalized proportion of the j-th type of outlier in the i1-th split block, and ; The value of the j-th type of anomaly index for the i1th split block; is the gray resolution coefficient, with a value of 0.1 to 0.5; k is the segmentation block number, with a value of 1, 2, or 3; In this embodiment, A value of 0.5 is recommended to adjust the resolution of gray relational degree; the smaller the value, the higher the resolution.

[0069] Anomaly based on composite architecture Construct an augmented mapping matrix for abnormal parameters, and determine the diagonal matrix of the split block structure and hydrodynamic added parameters. : ,in, For parameter correction coefficients; These are the radial span, circumferential arc length, and axial height of the protective skirt, respectively. With the minimization of composite anomaly degree, optimal matching of added parameters, maximization of interlocking structural stiffness, and optimal continuity of hydrodynamic field as the core objectives, a new objective function with multiple constraints is constructed. : ,in, , , It is a multi-objective balance coefficient, and it belongs to the range of 0 to 1; Let be the stiffness of the interlocking structure of the i1th split block, and ; Let be the hydrodynamic continuity coefficient of the i1th split block, and ; ,and This represents the maximum permissible value for the anomaly degree of the composite architecture. This is the baseline value for the radial span of the protective skirt; The design reference stiffness for the interlocking structure; Minimum allowable stiffness; This represents the maximum permissible value for the anomaly degree of the composite architecture.

[0070] In this embodiment, .

[0071] Preferably, after constructing the multi-constraint objective function, it also includes: Based on the newly added objective function with multiple constraints, an adaptive quasi-Newton iterative optimization formula based on augmented Lagrange is used to perform co-optimization of the architecture and parameters for each split block: ,in, For the i1th split block in the tth iteration, the structure-hydrodynamic composite parameter vector is used. For adaptive iteration step size; To add the first-order partial derivative of the objective function; The Hessian matrix is ​​added as a new objective function; These are the regularization coefficients of the Hessian matrix; It is the identity matrix; Let be the Lagrange multiplier for the c-th constraint; Let c be the c-th constraint function; n is the total number of constraints. For the i1th split block in the (t+1)th iteration, the structure-hydrodynamic composite parameter vector is used. Iterate to When all constraints are met, a qualified three-stage interlock module is obtained, wherein, The iteration threshold; Based on the optimal parameter combination of the simulation qualified module, the automatic and precise disassembly of the protective skirt is completed through bidirectional mapping of geometric parameters and disassembly topology.

[0072] In this embodiment, the Lagrange multiplier iterative update formula is as follows: ,and For the c-th constraint of the i1th split block in the (t+1)th iteration; The step size for updating the Lagrange multipliers, ranging from 0.01 to 0.1, is related to the iteration step size. Matching adjustments.

[0073] In this embodiment, the modular disassembly of the monopile wind turbine protective skirt only adopts a geometrically equal mechanical disassembly method, which only verifies the geometric dimensions of the module and does not consider the multi-physics field coupling effects of hydrodynamic distortion, structural stress, and seabed interaction. After disassembly, the module is prone to problems such as excessive shear stress on the bed surface, stress concentration on the interlocking surface, and disruption of water flow continuity. Moreover, anomaly detection relies on manual on-site measurement and lacks quantitative evaluation and proactive optimization logic. This invention adopts an immersion boundary method-morphodynamic coupling multi-physics field full-domain simulation verification mechanism before disassembly, combined with a customized multi-physics field comprehensive anomaly quantification formula, improved information entropy-grey relational degree composite anomaly evaluation, and augmented Lagrange adaptive quasi-Newton iterative optimization, forming a closed-loop logic of pre-disassembly prediction-anomaly quantification-root cause location-parameter optimization-automatic disassembly. This is an improvement on the Gaussian sinusoidal key interlocking module of the monopile biomimetic scour protection system.

[0074] All calculations in this invention are based on the 36 sets of fixed design variable combinations preset by the Taguchi L36 orthogonal array in the original technical solution. These 36 sets of samples have completely covered the value range of all core design variables such as the diameter ratio of the protective skirt, the parameters of the protective pin, and the area ratio. There is no need to collect additional real-time samples on site or to dynamically expand the sample library. The calculation process only needs to call the pre-made 36 sets of orthogonal sample data. There is no technical defect of relying on a large number of samples. The sample size is fixed and can be pre-made and stored.

[0075] This invention performs dimensionality reduction and parameter lightweighting for multiphysics calculations, simplifying the original 4 types of physical field deviations and 12 calculation parameters into 4 core indicators: bed shear stress, structural stress, geometric tolerance, and hydrodynamic continuity. Furthermore, parameters such as the analytic hierarchy process-entropy weight composite weight, gray resolution coefficient, and parameter correction coefficient are all pre-defined engineering values ​​(e.g., the gray resolution coefficient is fixed at 0.5, and the parameter correction coefficient is fixed at 0.8), eliminating the need for real-time iterative weight calculation. At the same time, the global quantization formula, composite anomaly formula, and iterative optimization formula are compiled into an embedded one-click calculation template, which is integrated into the lightweight software of the offshore wind power operation and maintenance terminal. The calculation process can be completed in only 10-15 seconds, without any complex manual calculation steps.

[0076] This technical solution does not require real-time online execution and perfectly matches the periodic patterns of offshore operation and maintenance of monopile wind turbines. The specific execution cycle is as follows: During the factory prefabrication stage: perform one split optimization calculation to determine the initial split parameters of the module; During the offshore operation and maintenance stage: perform one verification calculation according to the tidal cycle (every 15 days) or the quarterly inspection cycle. The calculation only needs to be completed by calling the prefabricated template through the terminal during the operation and maintenance window, without the need for continuous real-time operation. This execution frequency is fully adapted to the operation and maintenance rhythm of offshore wind power projects, and the calculation time and execution conditions meet the requirements of offshore on-site operation.

[0077] In this embodiment, the analytic hierarchy process (AHP)-entropy weight composite weights are: w1=0.35 (bed shear stress), w2=0.30 (structural stress), w3=0.20 (geometric tolerance), and w4=0.15 (hydrodynamic continuity), satisfying w1+w2+w3+w4=1. It should be noted that w1 can be adjusted according to the sea conditions and geological conditions of the project, w2 can be adjusted according to the material properties and load conditions, w3 can be adjusted according to the manufacturing and assembly precision requirements, and w4 can be adjusted according to the sea current conditions.

[0078] Preset threshold for acceptable abnormal indicators: =0.15, which is based on simulation verification of 100 splitting schemes. A value lower than this can guarantee that the performance degradation after splitting is <5%; Iteration threshold: =0.10, which is based on 50 sets of iterative optimization verification. A value lower than this can guarantee that the splitting scheme meets all constraints. Parameter correction factor: =0.8, iterative convergence verification; this value avoids iterative divergence while ensuring optimization efficiency; multi-objective balance coefficient: =0.25、 =0.35、 =0.40.

[0079] Gray resolution coefficient: =0.5, which is a commonly used value in grey system theory to ensure the resolution and stability of anomaly calculation; the grid convergence index (GCI) qualification standard is: GCI≤1.0%; Constant refinement ratio r=2; confidence factor =1.25.

[0080] Adaptive iteration step size The initial value was 0.15, which decreased to 0.05 after 10 iterations. The maximum number of iterations was 50. Verification showed that more than 95% of the solutions could converge to the threshold requirement within 30 iterations.

[0081] In this embodiment, the high-performance fiber-reinforced concrete (HPFRC) material mix proportion is as follows: mass ratio: cement grade 42.5 Quartz sand , steel fiber (Length 13mm, Diameter 0.2mm), fly ash Water-reducing agent Water-to-binder ratio: 0.32.

[0082] In this embodiment, ; ; ; .

[0083] In this embodiment, m=4 represents 4 types of abnormal indicators.

[0084] In this embodiment, the iteration parameter is: adaptive step size. =0.05~0.15; Regularization coefficient =0.01; identity matrix I; Lagrange multipliers =0.1~0.3; Number of constraints n=4; Termination rule: The number of iterations is ≥50 and If ≤0.10, all constraints are met, and iteration stops.

[0085] In this embodiment, HPFRC material is used to fabricate a three-section interlocking module according to the following process: Ingredient mixing: Weigh the raw materials according to the proportions and force mix for 180 seconds to ensure uniform distribution of steel fibers; Two-stage volumetric casting: The first stage involves casting 70% of the slurry and compacting it with 50Hz vibration for 60 seconds; the second stage involves casting 30% of the slurry and compacting it with 50Hz vibration for 60 seconds (total compaction time 120 seconds). Curing process: 24h room temperature (20±2℃) curing → 48h immersion in water (water temperature 20±2℃) hydration → natural curing for 28 days; Accuracy testing: Geometric dimensional tolerance ±0.5mm, interlocking surface roughness Ra3.2, compressive strength ≥80MPa, flexural strength ≥12MPa.

[0086] The beneficial effects of the above technical solution are: it eliminates problems such as hydrodynamic distortion, excessive structural stress, and out-of-tolerance geometric tolerance of the split blocks in advance, so that the shear stress ratio of the bed surface after the three-section interlocking module is assembled can be stably maintained at 1.00 in a hydrodynamically neutral state, and the scour protection efficiency can be maintained at 38.8% without attenuation. Through adaptive quasi-Newton iterative optimization, the multi-constraint objective function can be solved quickly, and the abnormal problems of the split blocks can be accurately eliminated. Combined with bidirectional mapping, the automatic and accurate splitting of the protective skirt can be achieved, improving the efficiency and accuracy of modular splitting, and ensuring that the three-section interlocking module can meet the installation and protection requirements of the biomimetic scour protection system.

[0087] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A novel biomimetic scour protection system for monopile wind turbines, characterized in that, include: Using the Fibonacci spiral of sunflower seeds as a biomimetic prototype, a protective skirt is installed around the single pile of a monopile wind turbine foundation. Protective pins arranged in an interlaced Fibonacci spiral pattern are placed on the protective skirt to construct the initial configuration, which is regarded as a new biomimetic scour protection system for monopile wind turbines.

2. An optimization method for a novel biomimetic scour protection system for monopile wind turbines, characterized in that, Based on the novel biomimetic scour protection system for monopile wind turbines as described in claim 1, the following steps are performed, specifically including: Step 1: Determine the core design variables and set the value range for each design variable; Step 2: Experimental design was carried out using an orthogonal array to construct a three-dimensional CFD simulation model that couples the immersion boundary method with the bed sand morphology and dynamics model, and to perform single-pile scour protection simulation calculations on the combination of design variables. Step 3: Based on the simulation calculation results, conduct correlation matrix analysis, regression analysis and response optimization to determine the optimal combination of design variables for the novel biomimetic scour protection system for monopile wind turbines based on the pre-set optimization objectives; Step 4: The protective skirt corresponding to the optimal design variable combination is divided into three interlocked modules. The modules are spliced ​​using a Gaussian sine key radial interlock structure to obtain the final optimized monopile wind turbine biomimetic scour protection system.

3. The optimization method according to claim 2, characterized in that, Mesh independence was verified using the mesh convergence index (GCI), and a medium mesh size was selected for simulation calculations. The grid convergence index (GCI) is calculated as follows: ； ； Where p is the order of convergence; r is the constant refinement ratio; It is the convergence result of the scouring depth of the fine mesh; This is the convergence result of the scour depth in the medium grid; This is the convergence result of the scour depth of the coarse mesh; This is relative error; is the confidence factor.

4. The optimization method according to claim 2, characterized in that, The protective skirt corresponding to the optimal combination of design variables is divided into three interlocking modules, and the three interlocking modules are three 120° equal parts.

5. The optimization method according to claim 2, characterized in that, The final optimized monopile wind turbine biomimetic scour protection system is made of high-performance fiber-reinforced concrete. The modular components are prepared through a two-stage volume casting process, 50Hz vibration compaction for 120 seconds, 24-hour ambient temperature curing, and 48-hour immersion hydration.

6. The optimization method according to claim 2, characterized in that, Before assembling modules based on the optimal combination of design variables, the following steps are also included: The protective skirt corresponding to the optimal design variable combination is decomposed, and a global quantitative formula for the multi-physics integrated anomaly index of the decomposed blocks is established. : ,in, The multiphysics comprehensive anomaly index for the i1th split block; The relative deviation of the shear stress on the bed surface surrounding the i1th split block; The critical starting shear stress of the bed sand; The stress deviation of the Gaussian sinusoidal interlocking surface structure of the i-th split block; To achieve the allowable stress of high-performance fiber-reinforced concrete; The geometric profile tolerance deviation of the i1th split block; Tolerances are allowed for modular geometry; The deviation of the hydrodynamic continuity distortion coefficient for the i1th split block; To design the hydrodynamic continuity factor; , , , The weighting coefficients of the analytic hierarchy process (AHP)-entropy weight composite system satisfy the following conditions: ; Traverse the three split blocks If all split blocks If the preset splitting scheme is deemed acceptable, then... The default threshold for abnormal indicators is set. Otherwise, extract the multiphysics simulation anomaly architecture for each split block.

7. The optimization method according to claim 6, characterized in that, After extracting the simulation anomaly architecture for each split block, the following is also included: Construct a set of topological anomaly feature vectors and determine the anomaly degree of the improved information entropy-grey relational degree composite architecture: ,in, denoted as the composite architecture anomaly degree of the i1th split block; m is the total number of anomaly types in the topological anomaly feature vector set; Let be the normalized proportion of the j-th type of outlier in the i1-th split block, and ; The value of the j-th type of anomaly index for the i1th split block; is the gray resolution coefficient, with a value of 0.1 to 0.5; k is the segmentation block number, with a value of 1, 2, or 3; Anomaly based on composite architecture Construct an augmented mapping matrix for abnormal parameters, and determine the diagonal matrix of the split block structure and hydrodynamic added parameters. : ,in, For parameter correction coefficients; These are the radial span, circumferential arc length, and axial height of the protective skirt, respectively. With the minimization of composite anomaly degree, optimal matching of added parameters, maximization of interlocking structural stiffness, and optimal continuity of hydrodynamic field as the core objectives, a new objective function with multiple constraints is constructed. : ,in, , , It is a multi-objective balance coefficient, and it belongs to the range of 0 to 1; Let be the stiffness of the interlocking structure of the i1th split block, and ; Let be the hydrodynamic continuity coefficient of the i1th split block, and ; ,and This represents the maximum permissible value for the anomaly degree of the composite architecture. This is the baseline value for the radial span of the protective skirt; The design reference stiffness for the interlocking structure; Minimum allowable stiffness; This represents the maximum permissible value for the anomaly degree of the composite architecture.

8. The optimization method according to claim 7, characterized in that, After constructing the multi-constraint objective function, it also includes: Based on the newly added objective function with multiple constraints, perform co-optimization of architecture and parameters for each split block: ,in, For the i1th split block in the tth iteration, the structure-hydrodynamic composite parameter vector is used. For adaptive iteration step size; To add the first-order partial derivative of the objective function; The Hessian matrix is ​​added as a new objective function; These are the regularization coefficients of the Hessian matrix; It is the identity matrix; Let be the Lagrange multiplier for the c-th constraint; Let c be the c-th constraint function; n is the total number of constraints. For the i1th split block in the (t+1)th iteration, the structure-hydrodynamic composite parameter vector is used. Iterate to When all constraints are met, a qualified three-stage interlock module is obtained, wherein, The iteration threshold; Based on the optimal parameter combination of the simulation qualified module, the automatic and precise disassembly of the protective skirt is completed through bidirectional mapping of geometric parameters and disassembly topology.

Images