Multi-objective intelligent optimization design method, system and storage medium for the aerodynamic shape of blades in a wind tunnel scale model of a large wind turbine in shutdown state

Abstract

The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade of the present invention designs a wind turbine model so that the lift-to-drag ratio (low Reynolds number) of the airfoil of the wind turbine model under different wind speeds and different angle of attack ranges is close to that of the original full-scale model airfoil, and the lift-to-drag coefficient of the designed airfoil is solved by the Viterna method; the method of combining blade element theory with a multi-objective genetic algorithm is used to solve the accurate evaluation of the wind load of the wind tunnel test of the wind turbine with a large pitch angle under extreme wind conditions, so as to ensure that the overall lift and drag coefficient of the wind turbine model blade are consistent with the prototype, thereby ensuring that the aerodynamic load of the wind turbine blade in the wind tunnel experiment is consistent with the actual structure; finally, the wind load of the wind turbine blade can be accurately evaluated through the wind tunnel test, providing a theoretical basis for accurately evaluating the wind load of the wind turbine blade in actual engineering. The present invention also discloses a multi-objective intelligent optimization design system and storage medium for the aerodynamic shape of a wind tunnel scale model blade.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind load testing, and specifically relates to a multi-objective intelligent optimization design method, system and storage medium for the aerodynamic shape of a wind tunnel scale model blade under the action of a typhoon and a large wind turbine shutdown state. Background Art

[0002] Energy issues have become one of the most concerned topics, and many scholars are discussing and researching traditional energy and new renewable energy. Wind energy, as a new energy source that is renewable, occupies a small area, is pollution-free, and has relatively mature development and utilization technology, has been widely used and developed around the world.

[0003] In view of the complex sea conditions in my country's sea areas, in-depth research on the dynamic response of large offshore wind turbine systems under extreme wind conditions such as typhoons has important research significance and engineering application value for the design of offshore wind turbine systems and long-term safe and stable operation. It is crucial to accurately assess the offshore environmental loads on offshore wind turbines. Therefore, wind tunnel test research on the wind loads on offshore wind turbines under typhoons can provide key information for understanding and optimizing wind turbine design and operation. In the extreme strong wind wind tunnel test of offshore wind turbines, due to the Reynolds effect, the geometric scale model has a large gap with the actual wind turbine model in terms of drag and lift coefficients, which in turn affects the accuracy of wind load assessment. Therefore, in the wind tunnel test, how to achieve the consistency between the lift and drag coefficients of the scaled model blades and the lift and drag coefficients of the full-scale blades under extreme wind conditions when the wind turbine is shut down (or fails) is a difficult problem. Summary of the invention

[0004] In view of this, in order to address the deficiencies in the prior art, the present invention aims to provide a multi-objective intelligent optimization design method, system and storage medium for a wind tunnel scale model blade aerodynamic shape.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] The present invention first proposes a multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade, comprising the following steps:

[0007] Step 1: According to the similarity criterion of wind tunnel test and considering the cross-sectional size of the wind tunnel and the maximum wind speed of the wind tunnel, determine the scale ratio of the model and the wind speed scale ratio, and determine the wind speed range and the Reynolds number range of the model;

[0008] Step 2: Based on the lift-drag coefficient of the airfoil at the maximum lift-drag ratio position under the full-scale model at the actual wind speed, design a wind turbine model so that the lift-drag ratio of the designed airfoil of the wind turbine model at different wind speeds and different angles of attack is close to the original lift-drag ratio;

[0009] Step 3: Under different Reynolds numbers, calculate the lift and drag coefficient of the designed airfoil in the attack angle range of -15° to 20°, and use the Viterna method to extrapolate the attack angle range of the lift and drag coefficient curve to -180° to 180°;

[0010] Step 4: Use the blade element theory to calculate the overall lift and drag coefficient of the original full-scale model blade at an attack angle of -180° to 180°;

[0011] Step 5: Digitally model the wind turbine blades, use the chord length distribution function and the torsion angle distribution as optimization variables, and establish a multi-objective optimization model;

[0012] Step 6: Combine blade element theory and multi-objective genetic algorithm to optimize the chord length distribution and torsion angle distribution of the scaled model blade, so that the overall lift-drag coefficient of the optimized scaled model is consistent with the overall lift-drag coefficient of the original full-scale model blade at an attack angle of 0° to 90°.

[0013] Furthermore, in step 1, the wind turbine wind tunnel test scale similarity criterion is:

[0014] Determine the wind speed scale ratio based on the maximum wind speed capability of the wind tunnel:

[0015]

[0016] Where: V is the wind speed scale ratio; U MODEL is the wind speed of the model wind tunnel test; U FULL is the actual wind speed; n is the wind speed scale ratio;

[0017] The geometric scale ratio is determined based on the size of the wind tunnel and the blockage ratio:

[0018]

[0019] Where: L is the geometric scale ratio; D model is the model size; D full is the actual structural size; m is the scale ratio of the model;

[0020] The air density ratio is determined by making the air density of the conventional atmospheric boundary layer wind equal to the air density in the low wind speed wind tunnel test:

[0021]

[0022] Where: ρ is the air density ratio; ρ model is the air density in the wind tunnel test, ρ full is the actual structural wind ambient air density;

[0023] The wind load scale of the rigid model is derived according to the dimensional analysis method:

[0024]

[0025]

[0026] Where: F is the scale ratio of force; M is the scale ratio of the moment; F model is the force on the model; F full is the actual structural force; M model is the model moment; is the actual structural moment.

[0027] Furthermore, in step 3, the method of using the Viterna method to extrapolate the attack angle range of the lift-drag coefficient curve to -180° to 180° is:

[0028]

[0029] C D =B1sin 2 α+B2cosα

[0030] in,

[0031] c; 1.11+0.018AR

[0032]

[0033]

[0034]

[0035]

[0036] Where: C L is the lift coefficient; C D is the drag coefficient; represents the maximum drag coefficient; is the drag coefficient at stall; is the lift coefficient at stall; AR is the aspect ratio; α is the wind attack angle; α stall Indicates the wind angle of attack at stall.

[0037] Furthermore, in step 4, the method for calculating the overall lift-drag coefficient of the original full-scale model blade at an attack angle of -180° to 180° by blade element theory is:

[0038] Cut a blade element with a thickness of δr along the span direction of the fan blade. The blade element δr moves along the rotation plane, generating a reverse relative velocity. The wind speed passing through the fan will produce a velocity reduction, and the sum velocity of the wake on the fan blade element is:

[0039]

[0040]

[0041] Where: U T is the sum speed of the wake on the fan blade element; U0 is the axial wind speed; r is the radial position of the blade element; Ω is the angular velocity of the fan rotation; a is the axial induction factor; a' is the tangential induction factor; φ is the angle between the relative resultant velocity and the rotating plane; W is the sum speed; Ωr(1+a′) is the tangential induced velocity; U0(1-a) is the axial induced velocity;

[0042] From the velocity triangle, we can know that the relative resultant velocity and the corresponding trigonometric relationship are:

[0043] α=φ-β

[0044] Where: α is the angle of attack relative to the resultant velocity; β is the blade element twist angle;

[0045] Lift and U T The direction of resistance is at right angles to U T direction; Assuming that the lift and drag of any blade element are not disturbed by other blade elements, the lift and drag on any blade element are:

[0046]

[0047]

[0048] Where: δL is the lift of any blade element, δD is the drag of any blade element, C L (α) is the lift coefficient of any blade element at an attack angle of α; C D (α) is the drag coefficient at an attack angle of α; c is the blade chord length; ρ is the air density; δr is the blade element span;

[0049] Integrate the lift δL and drag δD of any blade element of the entire wind turbine blade along the blade span r to obtain the overall lift L and drag D of the entire blade:

[0050]

[0051]

[0052] The overall lift coefficient and overall drag coefficient of the blade are defined as follows:

[0053]

[0054]

[0055] Where: is the overall lift coefficient of the blade; is the overall drag coefficient of the blade; c T is the equivalent chord length of the entire wind turbine blade; and:

[0056]

[0057] Furthermore, in step 5, the steps of creating the multi-objective optimization model are:

[0058] 51) Design optimization variables:

[0059] Chord length c(r) and twist angle θ(r) are selected as optimization variables, and the chord length distribution is calculated using the prototype blade chord length distribution function c p (mr) is described by c at radial position r. p (mr) is scaled by the optimization parameter X(1) times to obtain the scaled blade chord length distribution function c(r):

[0060] c(r)=X(1)c p (mr) / m,0≤r≤R

[0061] The distribution curve of the blade twist angle θ(r) of the scaled model uses a quadratic polynomial function to perform least square fitting on the prototype blade twist angle, and the initial fitting parameters are used as the optimization starting values:

[0062] c=X(2)r 2 +X(3)r+X(4),0≤r≤R

[0063] Where: m is the scale ratio of the model; R is the rotor radius; X(2), X(3) and X(4) are the fitting parameters of the distribution curve of the blade twist angle θ(r);

[0064] 52) Determine the optimization objective function: Taking the minimum value of the overall lift coefficient and the overall drag coefficient of the blade at the design attack angle under different wind speeds as the objective function, the optimization objective function obtained is:

[0065]

[0066]

[0067] Where: f1(θ, c) and f2(θ, c) are the optimization objective functions of the overall lift coefficient and the overall drag coefficient respectively; U L and U UOptimize the minimum and maximum values ​​of the incoming wind speed corresponding to the working conditions respectively; C L (U) and are the overall lift coefficients of the scaled model and the original model blade respectively; C D (U) and are the overall resistance coefficients of the scaled model and the original model blades respectively; U is the incoming wind speed;

[0068] Convert dual-objective optimization into single-objective optimization:

[0069] f(θ,c)=λ1f1(θ,c)+λ2f2(θ,c)

[0070] Where: λ1 and λ2 represent weight coefficients, λ1,λ2∈[0,1] and λ1+λ2=1;

[0071] 53) Based on the optimization initial point as the fitting parameters of the chord length and the torsion angle, and setting a reasonable range to ensure that the model blade is as similar as possible to the geometrically similar blade, the constraint conditions are established:

[0072] θ L (r)≤θ(r)≤θ U (r)

[0073]

[0074] Where: θ L (r) and θ U (r) are the torsion angles at the blade tip and the blade root, respectively.

[0075] Further, in step 6, the principle of the multi-objective genetic algorithm is:

[0076] Objective function:

[0077] minF(x)=[f1(x),f2(x),...,f m (x)]

[0078] Constraints:

[0079] g i (x)≤0,i=1,2,...,k

[0080] h i (x)=0,i=1,2,...,l

[0081] a i ≤x i ≤ b i ,i=1,2,...,n

[0082] Where: f i (x) is the objective function; m is the number of objective functions; gi (x) is an inequality constraint; h i (x) is an equality constraint; x i are the design variables; k, l and n are the number of inequality constraints, equality constraints and design variables respectively.

[0083] The present invention also proposes a multi-objective intelligent optimization design system for the aerodynamic shape of a wind tunnel scale model blade, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described above when executing the program.

[0084] The present invention also proposes a storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0085] The beneficial effects of the present invention are:

[0086] The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade of the present invention designs a wind turbine model so that the lift-to-drag ratio (low Reynolds number) of the airfoil of the wind turbine model under different wind speeds and different angle of attack ranges is close to that of the original full-scale model airfoil, and the lift-drag coefficient of the designed airfoil is solved by the Viterna method; a method combining blade element theory and a multi-objective genetic algorithm is adopted to solve the accurate evaluation of wind loads in wind tunnel tests of wind turbines with large pitch angles under extreme wind conditions, so as to ensure that the overall lift and drag coefficients of the wind turbine model blades are consistent with those of the prototype, thereby ensuring that the aerodynamic loads of the wind turbine blades in the wind tunnel test are consistent with the actual structure; a wind turbine wind tunnel test optimization model is constructed based on the blade element theory and the multi-objective genetic algorithm, so that the overall lift-drag coefficients of the scaled model under low Reynolds numbers are consistent with the overall lift-drag coefficients of the original model under high Reynolds numbers; finally, the wind loads of wind turbine blades can be accurately evaluated through wind tunnel tests, providing a theoretical basis for accurately evaluating the wind loads of wind turbine blades in actual engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0088] Figure 1 It is a flow chart of the multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade of the present invention;

[0089] Figure 2 The full-scale model of a 15MW large wind turbine blade and the wind tunnel schematic diagram;

[0090] Figure 3 MATWIES6 airfoil diagram of wind tunnel test scale model designed for Profili software;

[0091] Figure 4 This is a comparison chart of lift and drag coefficients of the reference airfoil at full scale;

[0092] Figure 5 The lift and drag coefficient curves of the designed airfoil at different Reynolds numbers;

[0093] Figure 6 Fitting distribution diagram of quadratic function of blade twist angle;

[0094] Figure 7 This is a schematic diagram of the principle of blade element theory;

[0095] Figure 8 To optimize the blade chord length and twist angle distribution of the scaled model;

[0096] Fig. 9 Comparison of overall lift and drag coefficient of the optimized rear blade

[0097] Fig.10 Schematic diagram of the scaled model blade after optimization. DETAILED DESCRIPTION

[0098] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0099] like Figure 1 As shown, the multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade of this embodiment includes the following steps.

[0100] Step 1: According to the similarity criterion of wind tunnel test and considering the cross-sectional size of the wind tunnel and the maximum wind speed of the wind tunnel, determine the scale ratio of the model and the wind speed scale ratio, and determine the wind speed range and the Reynolds number range of the model.

[0101] Specifically, in this embodiment, the wind turbine wind tunnel test scale similarity criterion is:

[0102] (1) Since the test wind speed in the wind tunnel experiment is difficult to meet the Reynolds number requirements, the wind speed scale ratio is determined according to the maximum wind speed capacity of the wind tunnel:

[0103]

[0104] Where: V is the wind speed scale ratio; U MODEL is the wind speed of the model wind tunnel test; U FULL is the actual wind speed; n is the wind speed scale ratio.

[0105] (2) Determine the geometric scale ratio based on the size of the wind tunnel and the blockage ratio:

[0106]

[0107] Where: L is the geometric scale ratio; D model is the model size; D full is the actual structural size; m is the scale ratio of the model.

[0108] Specifically, wind tunnel blockage ratio = model downwind projection area / wind tunnel test section area; wind tunnel blockage ratio meets the requirement of ≤5%.

[0109] (3) The air density ratio is determined by assuming that the air density of the conventional atmospheric boundary layer wind is equal to the air density in the low wind speed wind tunnel test:

[0110]

[0111] Where: ρ is the air density ratio; ρ model is the air density in the wind tunnel test, ρ full is the actual structural wind ambient air density.

[0112] (4) After determining the above scale ratio, the wind load scale of the rigid model is derived according to the dimensional analysis method:

[0113]

[0114]

[0115] Where: F is the scale ratio of force; M is the scale ratio of the moment; F model is the force on the model; F full is the actual structural force; M model is the model moment; M full is the actual structural moment.

[0116] Step 2: Based on the lift-drag coefficient of the airfoil at the maximum lift-drag ratio position under the full-scale model under actual wind speed, use Profili software to design a wind turbine model so that the lift-drag ratio of the designed airfoil of the wind turbine model at different wind speeds and different angles of attack is close to that of the original full-scale model airfoil.

[0117] Step 3: Under different Reynolds numbers, the lift and drag coefficient of the designed airfoil in the attack angle range of -15° to 20° is calculated using Xfoil software, and the attack angle range of the lift and drag coefficient curve is extrapolated to -180° to 180° using AirfoilsPrep software and the Viterna method.

[0118] Specifically, in this embodiment, the method of using the Viterna method to extrapolate the attack angle range of the lift-drag coefficient curve to -180° to 180° is:

[0119]

[0120] C D =B1sin 2 α+B2cosα

[0121] in,

[0122] c; 1.11+0.018AR

[0123]

[0124] B1=C Dmax

[0125]

[0126]

[0127] Where: C L is the lift coefficient; C D is the drag coefficient; represents the maximum drag coefficient; is the drag coefficient at stall; is the lift coefficient at stall; AR is the aspect ratio; α is the wind attack angle; α stall Indicates the wind angle of attack at stall.

[0128] Step 4: Use the blade element theory to calculate the overall lift-drag coefficient of the original full-scale model blade at an attack angle of -180° to 180°.

[0129] Specifically, in this embodiment, the method for calculating the overall lift-drag coefficient of the original full-scale model blade at an attack angle of -180° to 180° by blade element theory is:

[0130] Cut a blade element with a thickness of δr along the span direction of the fan blade. The blade element δr moves along the rotation plane, generating a reverse relative velocity. The wind speed passing through the fan will produce a velocity reduction, and the sum velocity of the wake on the fan blade element is:

[0131]

[0132]

[0133] Where: U Tis the sum speed of the wake on the fan blade element; U0 is the axial wind speed; r is the radial position of the blade element; Ω is the angular velocity of the fan rotation; a is the axial induction factor; a' is the tangential induction factor; φ is the angle between the relative resultant velocity and the rotating plane; W is the sum speed; Ωr(1+a′) is the tangential induced velocity; U0(1-a) is the axial induced velocity;

[0134] From the velocity triangle, we can know that the relative resultant velocity and the corresponding trigonometric relationship are:

[0135] α=φ-β

[0136] Where: α is the angle of attack relative to the resultant velocity; β is the blade element twist angle.

[0137] Lift and U T The direction of resistance is at right angles to U T direction; Assuming that the lift and drag of any blade element are not disturbed by other blade elements, the lift and drag on any blade element are:

[0138]

[0139]

[0140] Where: δL is the lift of any blade element, δD is the drag of any blade element, C L (α) is the lift coefficient of any blade element at an attack angle of α; C D (α) is the drag coefficient at an attack angle of α; c is the blade chord length; ρ is the air density; δr is the blade element span;

[0141] Integrate the lift δL and drag δD of any blade element of the entire wind turbine blade along the blade span r to obtain the overall lift L and drag D of the entire blade:

[0142]

[0143]

[0144] The overall lift coefficient and overall drag coefficient of the blade are defined as follows:

[0145]

[0146]

[0147] Where: is the overall lift coefficient of the blade; is the overall drag coefficient of the blade; c T is the equivalent chord length of the entire wind turbine blade; and:

[0148]

[0149] Step 5: Digitally model the wind turbine blades. The mathematical model of the wind turbine blades is usually determined by the distribution functions of the airfoil, chord length and torsion angle along the span of the blade. In this embodiment, the chord length distribution function and the torsion angle distribution are used as optimization variables to establish a multi-objective optimization model.

[0150] Specifically, in this embodiment, the steps of the method for creating a multi-objective optimization model are as follows:

[0151] 51) Design optimization variables:

[0152] Chord length c(r) and twist angle θ(r) are selected as optimization variables, and the chord length distribution is calculated using the prototype blade chord length distribution function c p (mr) is described by c at radial position r. p (mr) is scaled by the optimization parameter X(1) times to obtain the scaled blade chord length distribution function c(r):

[0153] c(r)=X(1)c p (mr) / m,0≤r≤R

[0154] The distribution curve of the blade twist angle θ(r) of the scaled model uses a quadratic polynomial function to perform least square fitting on the prototype blade twist angle, and the initial fitting parameters are used as the optimization starting values:

[0155] c=X(2)r 2 +X(3)r+X(4),0≤r≤R

[0156] Where m is the model scale; R is the rotor radius; X(2), X(3) and X(4) are the fitting parameters of the distribution curve of the blade twist angle θ(r).

[0157] 52) Determine the optimization objective function: Taking the minimum value of the overall lift coefficient and the overall drag coefficient of the blade at the design attack angle under different wind speeds as the objective function, the optimization objective function obtained is:

[0158]

[0159]

[0160] Where: f1(θ, c) and f2(θ, c) are the optimization objective functions of the overall lift coefficient and the overall drag coefficient respectively; U L and U U Optimize the minimum and maximum values ​​of the incoming wind speed corresponding to the working conditions respectively; C L (U) and are the overall lift coefficients of the scaled model and the original model blade respectively; C D (U) and are the overall resistance coefficients of the scaled model and the original model blades respectively; U is the incoming wind speed.

[0161] Convert dual-objective optimization into single-objective optimization:

[0162] f(θ,c)=λ1f1(θ,c)+λ2f2(θ,c)

[0163] Where: λ1 and λ2 represent weight coefficients, λ1,λ2∈[0,1] and λ1+λ2=1.

[0164] 53) Based on the optimization initial point as the fitting parameters of the chord length and the torsion angle, and setting a reasonable range to ensure that the model blade is as similar as possible to the geometrically similar blade, the constraint conditions are established:

[0165] θ L (r)≤θ(r)≤θ U (r)

[0166]

[0167] Where: θ L (r) and θ U (r) are the torsion angles at the blade tip and the blade root, respectively.

[0168] Step 6: Combine blade element theory and multi-objective genetic algorithm to optimize the chord length distribution and torsion angle distribution of the scaled model blade, so that the overall lift-drag coefficient of the optimized scaled model is consistent with the overall lift-drag coefficient of the original full-scale model blade at an attack angle of 0° to 90°.

[0169] The essence of genetic algorithm is to use a global optimization probabilistic search algorithm to simulate the genetic, evolutionary and mutation processes of organisms under natural conditions. Solving multi-objective problems is to find a set of optimal solutions when given constraints and multiple objective function conditions are met. Specifically, in this embodiment, the principle of multi-objective genetic algorithm is:

[0170] Objective function:

[0171] minF(x)=[f1(x),f2(x),...,f m (x)]

[0172] Constraints:

[0173] g i (x)≤0,i=1,2,...,k

[0174] h i (x)=0,i=1,2,...,l

[0175] a i ≤x i ≤ bi ,i=1,2,...,n

[0176] Where: f i (x) is the objective function; m is the number of objective functions; g i (x) is an inequality constraint; h i (x) is an equality constraint; x i are the design variables; k, l and n are the number of inequality constraints, equality constraints and design variables respectively.

[0177] The implementation steps of the genetic algorithm are to group and number each possible variable in the solution space, and the numbered variables are chromosomes (individuals). Multiple possible solutions are grouped together, called the initial population, and the fitness of each solution is evaluated and calculated according to the predetermined objective function constraint value, so that the solution with better performance has a higher fitness. It is replicated, and a new generation of population is generated through the selection of genetic operators, crossover recombination, and mutation. In this way, the new generation of population is better adapted to the objective function than the previous generation through continuous selection and recombination from generation to generation, and finally the optimal solution or the optimal approximate solution to the problem is obtained.

[0178] The specific implementation of the method of this embodiment is described in detail below with reference to specific examples.

[0179] In this embodiment, the 15MW wind turbine model used is as follows: Figure 2 As shown in (a), the blade length is 120m and the height of the entire model is 270m. The cross-sectional dimensions of the wind tunnel are 1.8 (width) × 2.4 (height), as shown in Figure 2 As shown in (b), the wind speed range of the wind tunnel is 0-35m / s, and the model scale is determined by satisfying the blockage ratio.

[0180] The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade of this embodiment includes the following steps.

[0181] Step 1: Determine the scale ratio of the model and the wind speed scale ratio according to the similarity criterion of the wind tunnel test and consider the cross-sectional size of the wind tunnel and the maximum wind speed of the wind tunnel, and determine the wind speed range and the Reynolds number range of the model;

[0182] The wind speed scale ratio is determined to be 1:5 based on the maximum wind speed capability of the wind tunnel, namely:

[0183]

[0184] The geometric scale ratio is determined according to the size of the wind tunnel, the structural size, and the requirements of the blockage ratio. Wind tunnel blockage ratio = model downwind projection area / wind tunnel test section area; wind tunnel blockage ratio meets the requirement of ≤5%. The geometric scale ratio is as follows:

[0185]

[0186] The air density ratio is determined by equating the air density of the conventional atmospheric boundary layer wind with the air density in the low wind speed wind tunnel test:

[0187]

[0188] Step 2: Based on the lift-drag coefficient of the airfoil at the maximum lift-drag ratio position under the full-scale model at the actual wind speed, use Profili software to design an airfoil with a lift-drag ratio (low Reynolds number) close to the original lift-drag ratio under different wind speeds and different angles of attack of the wind turbine model;

[0189] According to the full-scale model blade of a 15MW wind turbine, the airfoil with the largest lift-to-drag ratio at high Reynolds number is selected as a design reference. Then, a new airfoil is designed using Profili software, and the lift-to-drag ratio is similar to that of the high Reynolds number airfoil selected by the original full-scale model at low Reynolds number. The designed MATWIES6 airfoil is as follows: Figure 3 The lift and drag coefficients of the designed MATWIES6 airfoil and the original full-scale model are shown in Figure 2. Figure 4 shown.

[0190] Step 3: Use Xfoil software to calculate the lift-drag coefficient of the designed MATWIES6 airfoil at different Reynolds numbers (low Reynolds number) in the attack angle range of -15° to 20°, and use AirfoilsPrepV2.0.3 software and Viterna method to expand the attack angle range of the lift-drag coefficient curve to -180° to 180°.

[0191] Specifically, according to the designed airfoil obtained in step 2, Xfoil is used to calculate the lift and drag coefficient of the designed airfoil at different Reynolds numbers (low Reynolds numbers) in the range of -15° to 20° attack angles, and then the attack angle range of the lift and drag coefficient curve is expanded to -180° to 180° according to the Viterna method. The lift and drag coefficients of the designed MATWIES6 airfoil at different Reynolds numbers (low Reynolds numbers) are as follows: Figure 5 The extended Viterna method for lift-drag coefficient is:

[0192]

[0193] C D =B1sin 2 α+B2cosα

[0194] in,

[0195] c≈1.11+0.018AR

[0196]

[0197]

[0198]

[0199]

[0200] In the formula, C L is the lift coefficient, C D is the drag coefficient, represents the maximum drag coefficient, is the drag coefficient at stall, is the lift coefficient at stall, AR is the aspect ratio, α is the wind attack angle, and α stall Indicates the wind angle of attack at stall.

[0201] The fourth step is to calculate the overall lift and drag coefficient of the original full-scale model blade at an attack angle of -180° to 180° using the blade element theory, such as Figure 6 shown.

[0202] Specifically, Figure 7 As shown in the figure, the blade element theory for calculating the overall lift and drag coefficient of the original full-scale model blade at an attack angle of -180° to 180° is:

[0203] Cut a blade element with a thickness of δr along the span direction of the fan blade, where r is the distance between the blade element and the wheel load center. Assume that the fan's rotational angular velocity is Ω, and the blade element micro-segment δr moves along the rotation plane, then a reverse relative velocity will be generated. The wind speed passing through the fan will produce a velocity reduction, where the axial induction factor a is defined, and the tangential induction factor is a′, then the total velocity of the wake on the fan blade element is U T , usually called the "total speed" or "effective speed" from the blade element perspective.

[0204]

[0205]

[0206] Where: U T is the sum speed of the wake on the fan blade element; U0 is the axial wind speed; r is the radial position of the blade element; Ω is the angular velocity of the fan rotation; a is the axial induction factor; a' is the tangential induction factor; φ is the angle between the relative resultant velocity and the rotating plane; W is the sum speed; Ωr(1+a′) is the tangential induced velocity; U0(1-a) is the axial induced velocity;

[0207] From the velocity triangle, we can know that the relative resultant velocity and the corresponding trigonometric relationship are:

[0208] α=φ-β

[0209] from Figure 7 It can be seen that the lift and U T The direction of resistance is at right angles to U T According to the assumption of blade element theory, that is, the lift and drag of any blade element are not affected by other blade elements, the lift and drag on any blade element are:

[0210]

[0211]

[0212] Where: δL is the lift of any blade element, δD is the drag of any blade element, C L (α) is the lift coefficient of any blade element at an attack angle of α; C D (α) is the drag coefficient at an attack angle of α; c is the blade chord length; ρ is the air density; δr is the blade element span;

[0213] The overall lift L and drag D of the entire blade can be obtained by integrating the lift δL and drag δD of any blade element of the entire wind turbine blade along the blade span r:

[0214]

[0215]

[0216] Therefore, the overall lift coefficient and drag coefficient of the blade can be defined as follows:

[0217]

[0218]

[0219] where c T It is defined as the equivalent chord length of the entire wind turbine blade and is expressed as follows:

[0220]

[0221] The fifth step is to digitally model the wind turbine blades. The mathematical model of the wind turbine blades is usually determined by the airfoil, chord length and torsion angle distribution function along the blade span, that is, the wind turbine blade is determined by the airfoil coordinate data, the chord length distribution function and the torsion angle distribution function. In this embodiment, the chord length distribution function and the torsion angle distribution are used as optimization variables to establish an optimization model.

[0222] Specifically, the chord length c(r) and the twist angle θ(r) are selected as optimization variables, and the chord length distribution can be expressed by the prototype blade chord length distribution function c p (mr) is described by c at radial position r. p(mr) is scaled by the optimization parameter X(1) times to obtain the scaled blade chord length distribution function c(r). The distribution curve of the blade torsion angle θ(r) of the scaled model uses a quadratic polynomial function to perform a least squares fit on the prototype blade torsion angle, and the initial fitting parameters are used as the optimization starting value. The torsion angle distribution fitting is as follows: Figure 6 As shown. The optimized parameters for this design are as follows

[0223] c(r)=X(1)c p (mr) / m,0≤r≤R

[0224] c=X(2)r 2 +X(3)r+X(4),0≤r≤R

[0225] Where R is the rotor radius.

[0226] Then determine the target optimization function, and use the overall lift and drag coefficient of the blade at the design angle of attack under different wind speeds as the target function. The target optimization function is as follows:

[0227]

[0228]

[0229] Where, U is the incoming wind speed; U L and U U Optimize the minimum and maximum values ​​of the incoming wind speed corresponding to the working conditions respectively; C L (U) and are the overall lift coefficients of the scaled model and the original model blade respectively; C D (U) and are the overall drag coefficients of the scaled model and the original model blade, respectively. Subsequently, the dual-objective optimization is transformed into a single-objective optimization:

[0230] f(θ,c)=λ1f1(θ,c)+λ2f2(θ,c)

[0231] Where λ1 and λ2 represent weight coefficients, λ1,λ2∈[0,1] and λ1+λ2=1.

[0232] Finally, the initial point of optimization is used as the fitting parameters of the chord length and the torsion angle, and a reasonable range is set to ensure that the model blade is as similar as possible to the geometrically similar blade. The following constraints can be established:

[0233] θ L (r)≤θ(r)≤θ U (r)

[0234]

[0235] In the formula, θ L(r) and θ U (r) are the torsion angles at the blade tip and the blade root, respectively.

[0236] In the sixth step, the blade element theory and multi-objective genetic algorithm are combined to optimize the chord length distribution and torsion angle distribution of the scaled model blade, so that the overall lift and drag coefficient of the optimized scaled model is consistent with the overall lift and drag coefficient of the original full-scale model blade at an attack angle of 0° to 90°. The blade chord length and torsion angle distribution of the scaled model of the wind tunnel test after optimization are shown in the figure below: Figure 8 As shown; the comparison results of the overall lift and drag coefficient of the blade after optimization are shown in Fig. 9 As shown; the optimized scaled blade model obtained by final modeling is as follows Fig.10 shown. Fig. 9 Middle,U R is the fan design wind speed (unit: m / s); U M Wind speed for wind tunnel test of scaled model of wind turbine (unit: m / s).

[0237] Specifically, the essence of genetic algorithm is a global optimization probabilistic search algorithm that simulates the genetic, evolutionary and mutation processes of organisms under natural conditions. Solving multi-objective problems is to find a set of optimal solutions while satisfying given constraints and multiple objective function conditions. Its mathematical expression is:

[0238] Objective function: minF(x)=[f1(x),f2(x),...,f m (x)]

[0239] Constraints:

[0240] Where f i (x) is the objective function, m is the number of objective functions, g i (x) is an inequality constraint, h i (x) is an equality constraint, x i is a design variable.

[0241] The implementation steps of the genetic algorithm are to group and number each possible variable in the solution space, and the numbered variables are chromosomes (individuals). Multiple possible solutions are grouped together, called the initial population, and the fitness of each solution is evaluated and calculated according to the predetermined objective function constraint value, so that the solution with better performance has a higher fitness. It is replicated, and a new generation of population is generated through the selection of genetic operators, crossover recombination, and mutation. In this way, the new generation of population is better adapted to the objective function than the previous generation through continuous selection and recombination from generation to generation, and finally the optimal solution or the optimal approximate solution to the problem is obtained.

[0242] This embodiment also proposes a multi-objective intelligent optimization design system for the aerodynamic shape of a wind tunnel scale model blade, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described above when executing the program.

[0243] This embodiment further proposes a storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0244] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade, characterized by: The steps include: Step 1: According to the similarity criterion of wind tunnel test and considering the cross-sectional size of the wind tunnel and the maximum wind speed of the wind tunnel, determine the scale ratio of the model and the wind speed scale ratio, and determine the wind speed range and the Reynolds number range of the model; Step 2: Based on the lift-drag coefficient of the airfoil at the maximum lift-drag ratio position under the full-scale model at the actual wind speed, design a wind turbine model so that the lift-drag ratio of the designed airfoil of the wind turbine model at different wind speeds and different angles of attack is close to the original lift-drag ratio; Step 3: Under different Reynolds numbers, calculate the lift and drag coefficient of the designed airfoil in the attack angle range of -15° to 20°, and use the Viterna method to extrapolate the attack angle range of the lift and drag coefficient curve to -180° to 180°; Step 4: Use the blade element theory to calculate the overall lift and drag coefficient of the original full-scale model blade at an attack angle of -180° to 180°; Step 5: Digitally model the wind turbine blades, use the chord length distribution function and the torsion angle distribution as optimization variables, and establish a multi-objective optimization model; The steps for creating a multi-objective optimization model are as follows: 51) Design optimization variables: Chord length c(r) and twist angle θ(r) are selected as optimization variables, and the chord length distribution is calculated using the prototype blade chord length distribution function c p (mr) is described by c at radial position r p (mr) is scaled by the optimization parameter X(1) times to obtain the scaled blade chord length distribution function c(r): c(r)=X(1)c p (mr) / m,0≤r≤R The distribution curve of the blade twist angle θ(r) of the scaled model uses a quadratic polynomial function to perform least square fitting on the prototype blade twist angle, and the initial fitting parameters are used as the optimization starting values: c=X(2)r 2 +X(3)r+X(4),0≤r≤R Where: m is the scale ratio of the model; R is the rotor radius; X(2), X(3) and X(4) are the fitting parameters of the distribution curve of the blade twist angle θ(r); 52) Determine the optimization objective function: Taking the minimum value of the overall lift coefficient and the overall drag coefficient of the blade at the design attack angle under different wind speeds as the objective function, the optimization objective function obtained is: Where: f1(θ, c) and f2(θ, c) are the optimization objective functions of the overall lift coefficient and the overall drag coefficient respectively; U L and U U Optimize the minimum and maximum values ​​of the incoming wind speed corresponding to the working conditions respectively; C L (U) and are the overall lift coefficients of the scaled model and the original model blade respectively; C D (U) and are the overall drag coefficients of the scaled model and the original model blades respectively; U is the incoming wind speed; Convert dual-objective optimization into single-objective optimization: f(θ,c)=λ1f1(θ,c)+λ2f2(θ,c) Where: λ1 and λ2 represent weight coefficients, λ1, λ2∈[0,1] and λ1+λ2=1; 53) Based on the optimization initial point as the fitting parameters of the chord length and the torsion angle, and setting a reasonable range to ensure that the model blade is as similar as possible to the geometrically similar blade, the constraint conditions are established: i L (r)≤θ(r)≤θ U (r) Where: θ L (r) and θ U (r) are the torsion angles at the blade tip and the blade root, respectively; Step 6: Combine blade element theory and multi-objective genetic algorithm to optimize the chord length distribution and torsion angle distribution of the scaled model blade, so that the overall lift-drag coefficient of the optimized scaled model is consistent with the overall lift-drag coefficient of the original full-scale model blade at an attack angle of 0° to 90°.

2. The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade according to claim 1 is characterized by: In step 1, the wind turbine wind tunnel test scale similarity criterion is: Determine the wind speed scale ratio based on the maximum wind speed capability of the wind tunnel: Where: V is the wind speed scale ratio; U MODEL is the wind speed of the model wind tunnel test; U FULL is the actual wind speed; n is the wind speed scale ratio; The geometric scale ratio is determined based on the size of the wind tunnel and the blockage ratio: Where: L is the geometric scale ratio; D model is the model size; D full is the actual structural size; m is the scale ratio of the model; The air density ratio is determined by making the air density of the conventional atmospheric boundary layer wind equal to the air density in the low wind speed wind tunnel test: Where: ρ is the air density ratio; ρ model is the air density in the wind tunnel test, ρ full is the actual structural wind ambient air density; The wind load scale of the rigid model is derived according to the dimensional analysis method: Where: F is the scale ratio of force; M is the scale ratio of the moment; F model is the force on the model; F full is the actual structural force; M model is the model moment; is the actual structural moment.

3. The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade according to claim 1, characterized in that: In step 3, the method of using the Viterna method to extrapolate the attack angle range of the lift-drag coefficient curve to -180° to 180° is: C D =B1sin 2 α+B2cosα in, Where: C L is the lift coefficient; CD is the drag coefficient; represents the maximum drag coefficient; is the drag coefficient at stall; is the lift coefficient at stall; AR is the aspect ratio; α is the wind attack angle; a stall Indicates the wind angle of attack at stall.

4. The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade according to claim 1, characterized in that: In step 4, the method for calculating the overall lift-drag coefficient of the original full-scale model blade at an attack angle of -180° to 180° by blade element theory is: Cut a blade element with a thickness of δr along the span direction of the fan blade. The blade element δr moves along the rotation plane, generating a reverse relative velocity. The wind speed passing through the fan will produce a velocity reduction, and the sum velocity of the wake on the fan blade element is: Where: U T is the sum speed of the wake on the fan blade element; U0 is the axial wind speed; r is the radial position of the blade element; Ω is the angular velocity of the fan rotation; a is the axial induction factor; a′ is the tangential induction factor; φ is the angle between the relative resultant velocity and the rotating plane; W is the sum speed; Ωr(1+a′) is the tangential induced velocity; U0(1-a) is the axial induced velocity; From the velocity triangle, we can know that the relative resultant velocity and the corresponding trigonometric relationship are: α=φ-β Where: α is the angle of attack relative to the resultant velocity; β is the blade element twist angle; Lift and U T The direction of resistance is at right angles to U T direction; Assuming that the lift and drag of any blade element are not disturbed by other blade elements, the lift and drag on any blade element are: Where: δL is the lift of any blade element, δD is the drag of any blade element, C L (α) is the lift coefficient of any blade element at an attack angle of α; CD(α) is the drag coefficient at an attack angle of α; c is the chord length of the blade element; ρ is the air density; δr is the blade element span; Integrate the lift δL and drag δD of any blade element of the entire wind turbine blade along the blade span r to obtain the overall lift L and drag D of the entire blade: The overall lift coefficient and overall drag coefficient of the blade are defined as follows: Where: is the overall lift coefficient of the blade; is the overall drag coefficient of the blade; c T is the equivalent chord length of the entire wind turbine blade; and:

5. The multi-objective intelligent optimization design method for the aerodynamic shape of a wind tunnel scale model blade according to claim 1, characterized in that: In step 6, the principle of the multi-objective genetic algorithm is: Objective function: min F(x)=[f1(x),f2(x),...,f m (x)] Constraints: gi(x)≤0, i=1, 2,...,k h i (x)=0,i=1,2,...,l a i ≤x i ≤b i ,i=1,2,...,n Where: f i (x) is the objective function; m is the number of objective functions; g i (x) is an inequality constraint; h i (x) is an equality constraint; x i are the design variables; k, l and n are the number of inequality constraints, equality constraints and design variables respectively.

6. A multi-objective intelligent optimization design system for the aerodynamic shape of a wind tunnel scale model blade, characterized by: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 5 is implemented.

7. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 5.

Images