Floating fan pool test model blade optimization design method and device based on NSGA-III algorithm, medium and program

Abstract

The invention discloses a floating fan pool test model blade optimization design method and device based on an NSGA-III algorithm, a medium and a program. The method comprises the steps that firstly, the aerodynamic coefficient of a real fan blade airfoil and the Reynolds number of a floating fan model blade are calculated; selecting an appropriate low-Reynolds-number fan blade airfoil to replace a model blade airfoil and calculating an aerodynamic coefficient; calculating a target thrust value and a torque value of the floating fan model under different test required wind speeds; performing linear fitting to obtain a linear function of a lift coefficient linear region of the real type and the model blade relative to the attack angle; calculating the final chord length of the model blade; and finally, on the basis of an NSGA-III algorithm, the rotor rotating speed and the blade pitch angle of the floating fan model at different wind speeds are optimized by taking the minimum relative error of the thrust and the torque as a target. According to the invention, the thrust and torque of the model fan are matched with the target thrust and torque under different wind speed scales, the test accuracy is improved, and resources are saved.

Description

Technical Field

[0001] The invention relates to the technical field of marine engineering model tests, and in particular to a floating wind turbine pool test model blade optimization design method, device, medium and program based on NSGA-III algorithm. Background Art

[0002] Due to the ongoing environmental crisis, renewable energy technologies that utilize wind energy, ocean currents, and waves have attracted much attention. Among them, floating wind turbines are a promising offshore wind energy collection technology. In order to evaluate the performance of floating offshore wind turbine systems, it is necessary to conduct wind and wave tank tests, and model tests should meet the geometric and kinematic similarities between the actual wind turbine and the model wind turbine. For floating offshore wind turbines, the platform as the lower part is mainly affected by inertia, while the wind turbine as the upper part is mainly affected by viscosity.

[0003] In a wind and wave pool, the aerodynamic part needs to meet the Reynolds similarity, while the hydrodynamic test is generally carried out using Froude similarity, which results in a smaller lift coefficient for the blades of the geometric scale model. Therefore, the aerodynamic performance of the wind turbine model is different from that of the actual wind turbine. This phenomenon is called the scale effect. Therefore, the Reynolds number of the wind turbine model is lower than that of the real wind turbine. For this reason, several model test techniques for floating offshore wind turbines have been proposed, such as drag discs, steady forces, geometric similarity, and real-time hybrid tests. However, existing methods can only achieve similarity in the interaction between the floating foundation and the upper wind turbine structure at certain wind speeds.

[0004] In order to make the thrust and torque of the model fan satisfy the Froude similarity relationship with the thrust and torque of the full-size fan in the whole wind speed range, this patent proposes an efficient floating fan pool test model blade optimization design method.

[0005] Technical Solution

[0006] In order to solve the problem that it is difficult to correctly simulate the interaction between the floating support platform and the wind turbine in the scaled model test of the floating wind turbine, the present invention proposes a floating wind turbine pool test model blade optimization design method, device, medium and program based on the NSGA-III algorithm. The method reconstructs the geometric features of the blades and adjusts the control strategy of the wind turbine to make the wind turbine model meet the thrust and torque similar to those of the actual wind turbine. The specific steps of the method are as follows:

[0007] Step S1: determining the scale ratio of the model based on the specific requirements of the scaled model test of the floating wind turbine wave tank;

[0008] Step S2: Calculate the aerodynamic coefficients of the actual fan blade airfoil, including the lift coefficient and the drag coefficient;

[0009] Step S3: calculating the Reynolds number of the floating wind turbine model blades;

[0010] Step S4: Based on the results of steps S2 and S3, a suitable low Reynolds number fan blade airfoil is selected to replace the model blade airfoil;

[0011] Step S5: Calculate the aerodynamic coefficients of the low Reynolds number airfoil selected in step S4, including the lift coefficient and the drag coefficient;

[0012] Step S6: Calculating the target thrust value and torque value of the floating wind turbine model at different test required wind speeds;

[0013] Step S7: based on the calculation results of S2 and S6, linear fitting is performed to obtain the linear function of the lift coefficient linear region of the real blade and the model blade relative to the angle of attack;

[0014] Step S8: Based on the result of step S7, the final chord length of the model blade is calculated by using the aerodynamic coefficients and the initial chord length of the model blade in combination with the lift matching principle;

[0015] Step S9: Based on the NSGA-III algorithm, with the goal of minimizing the relative error of thrust and torque, setting the constraint condition that the relative error of thrust and torque is less than 5%, the rotor speed and blade pitch angle of the floating wind turbine model at different wind speeds are optimized.

[0016] Furthermore, when selecting a suitable scale ratio in step S1, the various requirements specified in the test task book and the function and size limitations of the wind and wave pool itself should be comprehensively considered, and the similarity criterion and the requirements of economy and feasibility should also be fully considered.

[0017] Furthermore, the lift coefficients in steps S2 and S5 are calculated using the software Xfoil developed by MIT. After setting the parameters in the software, the lift and drag coefficients of the airfoil at small angles of attack are calculated according to the panel method and boundary layer theory.

[0018] Furthermore, the correlation between the Reynolds numbers of the model fan and the real fan in step S3 is as follows:

[0019]

[0020] In the formula, λ Re is the ratio coefficient of the Reynolds number between the real fan and the model fan, Re is the Reynolds number, V is the characteristic speed, c is the chord length of the fan blade airfoil, v is the kinematic viscosity coefficient of the air, λ L is the length scale ratio, λ V is the speed scale ratio. The subscripts F and M represent the real floating wind turbine and the floating wind turbine model, respectively.

[0021] Furthermore, in step S4, at the Reynolds number after the fan is scaled down, the closer the lift-to-drag ratio of the low Reynolds number replacement airfoil is to that of the actual fan, the better the similarity of thrust and torque between the model fan and the actual fan.

[0022] Furthermore, in step S6, the wind speed in the model and the actual calculation should satisfy the Froude similarity condition, that is:

[0023]

[0024] Where g is the acceleration due to gravity and L is the characteristic length.

[0025] When calculating the target thrust and torque for the floating wind turbine model test, the modified blade element momentum theory is used:

[0026]

[0027]

[0028] Where, dT is the annular unit thrust, dQ is the annular unit torque, ρ is the fluid density, V0 is the axial flow velocity, a is the axial induced velocity, ω is the wind wheel angular velocity, B is the number of blades, r is the distance from the blade section to the blade root, is the angle between the incoming flow velocity and the blade disk, where C n and C t Respectively meet:

[0029]

[0030]

[0031] In the formula, C l is the lift coefficient, C d is the resistance coefficient.

[0032] The thrust and torque of the wind turbine are used as external loads of the hydrodynamic system of the floating support platform and are converted into:

[0033]

[0034]

[0035] Where T is the thrust of the wind turbine and Q is the torque of the rotor.

[0036] Scaled model tests are usually conducted in fresh water, while prototype systems usually operate in sea water. Therefore, the results of the tank model test need to be corrected according to the density of water. The correction factor γ is the ratio of the density of sea water to that of fresh water, and is generally taken as γ = 1.025. At this time, the similarity relationship between thrust and torque is:

[0037]

[0038]

[0039] Furthermore, the linear fitting equation obtained in step S7 can be expressed as:

[0040] C l (α) = k·α + C l0

[0041] In the formula, C l is the lift coefficient, k is the slope of the linear region of the lift curve, α is the angle of attack of the blade, C l0 is the lift coefficient at zero angle of attack.

[0042] Furthermore, the step S8 first calculates the initial chord length of the fan model blade by geometric similarity theory, and then uses these aerodynamic coefficients and the initial chord length of the model blade to calculate the final chord length of the model blade by lift matching principle. The cross-sectional airfoil chord length calculation formula of the model blade is:

[0043] c M =c M0 ·λ L =(c F / λ L )·(k F / k M )

[0044] In the formula, c M is the chord length of the thrust similarity model fan blade airfoil, c M0 is the initial chord length of the model fan blade airfoil calculated based on geometric similarity, c F is the chord length of the real fan blade, k F and k M They are the slopes of the linear parts of the lift coefficient curves of the real floating wind turbine and the model floating wind turbine, respectively.

[0045] During the reconstruction of the fan blade, the cross-sectional airfoil twist angle of the model blade is consistent with that of the real blade, satisfying the following relationship:

[0046] β F =β M

[0047] Furthermore, the airfoil at the root of the blade is cylindrical, the lift coefficient is close to zero, and its chord does not affect the aerodynamic characteristics of the floating wind turbine model. Therefore, in order to ensure the lightweight of the model blade and facilitate its connection with the hub of the floating wind turbine model, the chord length of the airfoil at the root of the model blade needs to be manually adjusted.

[0048] The core idea of ​​the non-dominated sorting genetic algorithm (NSGA-III) used in the design of the present invention is to continuously evolve the population by simulating natural selection and genetic operations to produce a set of non-dominated solutions. It uses a set of predefined reference points that are evenly distributed in the objective space to guide and distribute the individuals in the population. These reference points ensure that the solutions are evenly distributed in the high-dimensional target space. NSGA-III uses a selection mechanism based on these reference points to select the next generation of populations. The optimization process specifically includes the following steps:

[0049] (1) Initialize the population: Randomly generate an initial population consisting of multiple individuals, each of which is represented by a set of decision variables.

[0050] (2) Non-dominated sorting: For each individual, the individuals in the population are non-dominated sorted according to their dominance relationship to obtain frontiers of different levels.

[0051] (3) Crowding distance calculation: Calculate the crowding distance of each individual. The crowding reflects the density of solutions around the individual.

[0052] (4) Selection operation: Based on the stratification, distance and crowding of individuals, a part of the solutions is selected to create the next generation population.

[0053] (5) Crossover and mutation operations: Use crossover and mutation in the genetic algorithm on the selected individuals to generate new individuals.

[0054] (6) Update the population: merge the newly generated individuals with the current population to produce the next generation of population.

[0055] (7) Termination condition judgment: The present invention aims to minimize the relative error of thrust and torque. During the optimization process, a constraint condition of the relative error of thrust and torque being less than 5% is set, and other precisions can also be used. Determine whether to terminate the algorithm based on the preset termination condition, otherwise return to non-dominated sorting. The multi-objective optimization settings are as follows:

[0056] minimize:F1=(T F-M -T M ) / T F-M ; F2=(Q F-M -Q M ) / Q F-M

[0057]

[0058] Where, T M and T F-M are the thrust and target thrust of the model fan, Q M and Q F-M are the torque and expected torque of the model fan, Ωi is the rotor speed at different wind speeds, θ i is the pitch angle at different wind speeds.

[0059] By continuously iterating the above steps, NSGA-III can gradually converge to a set of better non-dominated solutions, which are optimal under multiple objective functions while maintaining the diversity and uniformity of the solution set.

[0060] The beneficial effects of the present invention are:

[0061] In the floating wind turbine tank test based on Froude similarity, the difficulty in achieving Reynolds similarity leads to differences between the thrust and torque of the floating wind turbine model and the expected thrust and torque. Compared with the existing tank model test method, the newly proposed control strategy optimization method based on the third-generation non-dominated sorting genetic algorithm (NSGA-III) wind turbine model and the lift matching blade reconstruction method can achieve the matching of the model wind turbine thrust and torque with the target thrust and torque under different wind speed scales, thereby improving the accuracy of the test and saving resources.

[0062] The present invention achieves similarity in thrust and torque between the scaled model of the floating wind turbine and the actual wind turbine by reconstructing the geometric characteristics of the model blades (mainly including the blade chord length and the torsion angle) and adjusting the control strategy of the model wind turbine (mainly including the rotor speed and the blade pitch angle) using the NSGA-III algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a flow chart of the optimization design method for floating wind turbine model blades with similar thrust and torque proposed by the present invention;

[0064] Figure 2 The thrust and torque values ​​of the IEA15MW floating wind turbine calculated by AeroDyn are shown in the figure.

[0065] Figure 3 The lift coefficient curve of S7032 airfoil at different Reynolds numbers calculated by Xfiol;

[0066] Figure 4 is the chord length of the blade airfoil of the IEA15MW floating wind turbine model at different speed scale ratios;

[0067] Figure 5 This is a flow chart of the rotor speed and pitch angle optimization method for a floating wind turbine model based on the NSGA-III algorithm;

[0068] Figure 6 After optimization by NSGA-III algorithm The rotor speed and pitch angle of the IEA15MW floating wind turbine model at different wind speeds;

[0069] Figure 7 This is a comparison chart between the thrust and torque at different wind speeds after model optimization and the target thrust and torque. DETAILED DESCRIPTION

[0070] In order to make the technical problems solved by the present invention, the technical solutions adopted and the technical effects achieved clearer, the technical solutions of the embodiments of the present invention are further described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0071] The present invention discloses a method for optimizing the design of a floating wind turbine pool test model based on the NSGA-III algorithm, including the reconstruction of floating wind turbine model blades and optimization of the wind turbine control strategy. Figure 1 As shown, the specific steps are:

[0072] Step S1: determining the scale ratio of the model based on the specific requirements of the scaled model test of the floating wind turbine wave tank;

[0073] Step S2: Calculate the aerodynamic coefficients of the actual fan blade airfoil, including the lift coefficient and the drag coefficient;

[0074] Step S3: calculating the Reynolds number of the floating wind turbine model blades;

[0075] Step S4: Based on the results of steps S2 and S3, a suitable low Reynolds number fan blade airfoil is selected to replace the model blade airfoil;

[0076] Step S5: Calculate the aerodynamic coefficients of the low Reynolds number airfoil selected in step S4, including the lift coefficient and the drag coefficient;

[0077] Step S6: Calculating the target thrust value and torque value of the floating wind turbine model at different test required wind speeds;

[0078] Step S7: based on the calculation results of S2 and S6, linear fitting is performed to obtain the linear function of the lift coefficient linear region of the real blade and the model blade relative to the angle of attack;

[0079] Step S8: Based on the result of step S7, the final chord length of the model blade is calculated by using the aerodynamic coefficients and the initial chord length of the model blade in combination with the lift matching principle;

[0080] Step S9: Based on the NSGA-III algorithm, with the goal of minimizing the relative error of thrust and torque, setting the constraint condition that the relative error of thrust and torque is less than 5%, the rotor speed and blade pitch angle of the floating wind turbine model at different wind speeds are optimized.

[0081] When selecting a suitable scale ratio in step S1, the various requirements specified in the test task book and the function and size limitations of the wind and wave pool itself should be comprehensively considered, and the similarity criterion and the requirements of economy and feasibility should also be fully considered. L =80.

[0082] Furthermore, the lift coefficients in steps S2 and S5 are calculated using the software Xfoil developed by MIT. After setting the parameters in the software, the lift and drag coefficients of the airfoil at small angles of attack are calculated according to the panel method and boundary layer theory.

[0083] The correlation between the Reynolds numbers of the model fan and the real fan in step S3 is as follows:

[0084]

[0085] In the formula, λ Re is the ratio coefficient of the Reynolds number between the real fan and the model fan, Re is the Reynolds number, V is the characteristic speed, c is the chord length of the fan blade airfoil, v is the kinematic viscosity coefficient of the air, λ L is the length scale ratio, λ V is the speed scale ratio. The subscripts F and M represent the real floating wind turbine and the floating wind turbine model, respectively.

[0086] In step S4, at the Reynolds number after the fan is scaled down, the closer the lift-to-drag ratio of the low Reynolds number replacement airfoil is to that of the actual fan, the better the similarity of thrust and torque between the model fan and the actual fan. In this case, the SD7032 airfoil is selected as the replacement airfoil for the model blade;

[0087] In step S6, the wind speed in the model and the actual calculation should meet the Froude similarity condition, that is:

[0088]

[0089] Where g is the acceleration due to gravity and L is the characteristic length.

[0090] When calculating the target thrust and torque during the floating wind turbine model test in step S6, the modified blade element momentum theory is used:

[0091]

[0092]

[0093] Where, dT is the annular unit thrust, dQ is the annular unit torque, ρ is the fluid density, V0 is the axial flow velocity, a is the axial induced velocity, ω is the wind wheel angular velocity, B is the number of blades, r is the distance from the blade section to the blade root, is the angle between the incoming flow velocity and the blade disk, where C n and C t Respectively meet:

[0094]

[0095]

[0096] In the formula, C l is the lift coefficient, C d is the resistance coefficient.

[0097] The thrust and torque of the wind turbine are used as external loads of the hydrodynamic system of the floating support platform and are converted into:

[0098]

[0099]

[0100] Where T is the thrust of the wind turbine and Q is the torque of the rotor.

[0101] Scaled model tests are usually conducted in fresh water, while prototype systems usually operate in sea water. Therefore, the results of the tank model test need to be corrected according to the density of water. The correction factor γ is the ratio of the density of sea water to that of fresh water, and is generally taken as γ = 1.025. At this time, the similarity relationship between thrust and torque is:

[0102]

[0103]

[0104] Figure 2 The thrust and torque values ​​of the IEA15MW floating wind turbine calculated based on AeroDyn are in good agreement with the reference data published by the IEA.

[0105] The linear fitting formula obtained in step S7 can be expressed as:

[0106] C l (α) = k·α + C l0

[0107] Where k is the slope of the lift curve in the linear region, C l0 is the lift coefficient at zero angle of attack.

[0108] The step S8 first calculates the initial chord length of the fan model blade by using the geometric similarity theory, and then uses these aerodynamic coefficients and the initial chord length of the model blade to calculate the final chord length of the model blade through the lift matching principle. The calculation formula for the cross-sectional airfoil chord length of the model blade is:

[0109] c M =c M0 ·λ L =(c F / λ L )·(k F / k M )

[0110] In the formula, c M is the chord length of the thrust similarity model fan blade airfoil, c M0 is the initial chord length of the model fan blade airfoil calculated based on geometric similarity, c F is the chord length of the real fan blade, k F and k M They are the slopes of the linear parts of the lift coefficient curves of the real floating wind turbine and the model floating wind turbine, respectively.

[0111] During the reconstruction of the fan blade, the cross-sectional airfoil twist angle of the model blade is consistent with that of the real blade, satisfying the following relationship:

[0112] β F =β M

[0113] Furthermore, the airfoil at the root of the blade is cylindrical, the lift coefficient is close to zero, and its chord does not affect the aerodynamic characteristics of the floating wind turbine model. Therefore, in order to ensure the lightweight of the model blade and facilitate its connection with the hub of the floating wind turbine model, the chord length of the airfoil at the root of the model blade needs to be manually adjusted.

[0114] Figure 4 is the chord length of the blade airfoil of the IEA15MW floating wind turbine model at different speed scale ratios.

[0115] The core idea of ​​the non-dominated sorting genetic algorithm (NSGA-III) used in the design of this invention is to continuously evolve the population by simulating natural selection and genetic operations to produce a set of non-dominated solutions. It uses a set of predefined reference points that are evenly distributed in the objective space to guide and distribute the individuals in the population. These reference points ensure that the solutions are evenly distributed in the high-dimensional target space. NSGA-III uses a selection mechanism based on these reference points to select the next generation of population. Figure 5 As shown, the optimization process specifically includes the following steps:

[0116] Initialize the population: Randomly generate an initial population containing multiple individuals, each of which is represented by a set of decision variables.

[0117] Non-dominated sorting: For each individual, the individuals in the population are non-dominated sorted according to their dominance relationship to obtain different levels of frontiers.

[0118] Crowding distance calculation: Calculate the crowding distance of each individual. The crowding reflects the density of solutions around the individual.

[0119] Selection operation: Based on the stratification, distance and crowding of individuals, a part of the solution is selected to create the next generation population.

[0120] Crossover and mutation operations: Use crossover and mutation in the genetic algorithm on the selected individuals to generate new individuals.

[0121] Update the population: merge the newly generated individuals with the current population to produce the next generation of population.

[0122] Termination condition judgment: The present invention aims to minimize the relative error of thrust and torque. During the optimization process, a constraint condition of thrust and torque relative error less than 5% is set, and other precisions can also be used. Determine whether to terminate the algorithm according to the preset termination condition, otherwise return to non-dominated sorting. The multi-objective optimization settings are as follows:

[0123] minimize:F1=(T F-M -T M ) / T F-M ; F2=(Q F-M -Q M ) / Q F-M

[0124]

[0125] Where, T M and T F-M are the thrust and expected thrust of the model fan, Q M and Q F-M are the torque and expected torque of the model fan, Ω i is the rotor speed at different wind speeds, θ i is the pitch angle at different wind speeds.

[0126] By continuously iterating the above steps, NSGA-III can gradually converge to a set of better non-dominated solutions, which are optimal under multiple objective functions while maintaining the diversity and uniformity of the solution set.

[0127] Figure 6 After optimization by NSGA-III algorithm The rotor speed and pitch angle of the IEA15MW floating wind turbine model at different wind speeds.

[0128] Figure 7 This is a comparison chart of thrust and torque at different wind speeds after the IEA15MW floating wind turbine test model is optimized according to the method of the present invention and the target thrust and torque. Figure 7 It can be seen that by using the method of the present invention to reconstruct the floating wind turbine model blades and using the NSGA-III algorithm to adjust the rotor speed and pitch angle under different wind speeds, the thrust and torque values ​​in the floating wind turbine model test can be well consistent with the target thrust, verifying the feasibility of the method of the present invention.

[0129] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A floating wind turbine pool test model blade optimization design method based on NSGA-III algorithm, characterized by: The specific steps are as follows: Step 1: Determine the scale ratio of the model based on the specific requirements of the scaled model test of the floating wind turbine wave tank; Step 2: Calculate the aerodynamic coefficients of the actual fan blade airfoil; after setting the parameters in the software, calculate the lift and drag coefficients of the airfoil at a small angle of attack according to the panel method and boundary layer theory; Step 3: Calculate the Reynolds number of the floating wind turbine model blades; Step 4: Based on the results of steps 2 and 3, select an appropriate low Reynolds number fan blade airfoil to replace the model blade airfoil; Step 5: According to the method of step 2, the aerodynamic coefficients of the low Reynolds number airfoil selected in step 4 are calculated, including the lift coefficient and the drag coefficient; Step 6: Calculate the target thrust and torque values ​​of the floating wind turbine model at different test required wind speeds; Step 7: Based on the calculation results of 2 and 6, linear fitting is performed to obtain the linear function of the lift coefficient linear region of the real and model blades relative to the angle of attack; Step 8: Based on the results of step 7, the final chord length of the model blade is calculated using these aerodynamic coefficients and the initial chord length of the model blade in combination with the lift matching principle; Step 9: Based on the NSGA-III algorithm, with the goal of minimizing the relative error of thrust and torque, the constraint condition that the relative error of thrust and torque is less than 5% is set to optimize the rotor speed and blade pitch angle of the floating wind turbine model at different wind speeds.

2. The method for optimizing blade design of a floating wind turbine pool test model based on the NSGA-III algorithm according to claim 1, characterized in that: The step 3 calculates the Reynolds number correlation between the model and the actual floating wind turbine model blade: In the formula, λ Re is the ratio coefficient of the Reynolds number between the real fan and the model fan; Re is the Reynolds number; V is the characteristic speed; c is the chord length of the fan blade airfoil; v is the kinematic viscosity coefficient of the air; λ L is the length scale ratio; V is the speed scale ratio; subscripts F and M represent the real floating wind turbine and the model floating wind turbine, respectively.

3. The method for optimizing the blade design of a floating wind turbine pool test model based on the NSGA-III algorithm according to claim 1, characterized in that: In step 4, at the Reynolds number after the fan is scaled down, the closer the lift-to-drag ratio of the low Reynolds number replacement airfoil is to that of the actual fan, the better the similarity of thrust and torque between the model fan and the actual fan.

4. The method for optimizing the blade design of a floating wind turbine pool test model based on the NSGA-III algorithm according to claim 1, characterized in that: The step 6 is specifically as follows: The wind speed in the model and actual calculation should meet the Froude similarity condition, that is: Where g is the acceleration due to gravity and L is the characteristic length; When calculating the target thrust and torque of the floating wind turbine model test, the modified blade element momentum theory is used; the thrust and torque of the wind turbine are used as the external loads of the hydrodynamic system of the floating support platform, and are converted into: In the formula, T is the thrust of the wind turbine, and Q is the torque of the rotor; Scaled model tests are usually conducted in fresh water, while prototype systems usually operate in sea water; therefore, the results of the tank model test need to be corrected according to the density of water; the correction factor γ is the ratio of the density of sea water to that of fresh water, γ = 1.025; at this time, the similarity relationship between thrust and torque is:

5. The method for optimizing blade design of a floating wind turbine pool test model based on NSGA-III algorithm according to claim 1, characterized in that: In step 7, linear fitting is performed to obtain a linear function of the lift coefficient linear region of the real and model blades relative to the angle of attack, and the formula is expressed as follows: C l (a)=k·a+C l0 In the formula, C l is the lift coefficient, α is the angle of attack of the blade, k is the slope of the linear region of the lift curve, C l0 is the lift coefficient at zero angle of attack.

6. The method for optimizing blade design of a floating wind turbine pool test model based on NSGA-III algorithm according to claim 1, characterized in that: The step 8 is specifically as follows: The initial chord length of the fan model blade is calculated by geometric similarity theory: c M =c M0 ·l L =(c F / l L )·(k F / k M ) In the formula, c M is the chord length of the thrust similarity model fan blade airfoil, c M0 is the initial chord length of the model fan blade airfoil calculated based on geometric similarity, c F is the chord length of the real fan blade, k F and k M are the slopes of the linear parts of the lift coefficient curves of the real floating wind turbine and the model floating wind turbine, respectively; During the reconstruction of the fan blade, the cross-sectional airfoil twist angle of the model blade is consistent with that of the real blade, satisfying the following relationship: β F ＝β M The airfoil at the root of the blade is cylindrical, the lift coefficient is close to zero, and its chord does not affect the aerodynamic characteristics of the floating wind turbine model; to ensure the lightweight of the model blade and facilitate its connection with the hub of the floating wind turbine model, the chord length of the airfoil at the root of the model blade needs to be manually adjusted.

7. The method for optimizing blade design of a floating wind turbine pool test model based on NSGA-III algorithm according to claim 1, characterized in that: In step 9, the NSGA-III algorithm uses a set of predefined reference points, which are evenly distributed in the objective space, to guide and distribute the individuals in the population; these reference points ensure that the uniform distribution of the knowledge in the high-dimensional target space; NSGA-III uses a selection mechanism based on these reference points to select the next generation population; the optimization process specifically includes the following steps: (1) Initialize the population: randomly generate an initial population containing multiple individuals, each of which is represented by a set of decision variables; (2) Non-dominated sorting: For each individual, the individuals in the population are non-dominated sorted according to their dominance relationship to obtain frontiers of different levels; (3) Crowding distance calculation: Calculate the crowding distance of each individual. The crowding reflects the density of solutions around the individual. (4) Selection operation: select a part of the solution to create the next generation population based on the individual stratification, distance and crowding degree; (5) Crossover and mutation operations: Use crossover and mutation in the genetic algorithm on the selected individuals to generate new individuals; (6) Update the population: merge the newly generated individuals with the current population to produce the next generation of population; (7) Termination condition judgment: The goal is to minimize the relative error of thrust and torque. During the optimization process, a constraint condition of the relative error of thrust and torque less than 5% or other precision is set. The algorithm is terminated according to the preset termination condition, otherwise it returns to the non-dominated sorting. The multi-objective optimization settings are as follows: minimize:F1=(|T F-M -T M |) / T F-M ;F2=(|Q F-M -Q M |) / Q F-M Where, T M and T F-M are the thrust and target thrust of the model fan, Q M and Q F-M are the torque and expected torque of the model fan, Ω i is the rotor speed at different wind speeds, θ i is the pitch angle at different wind speeds.

8. A computer device / equipment / system comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images