Multi-round multi-target design method for deep sea water turbine

Abstract

The invention relates to the technical field of new energy. The invention discloses a multi-round multi-objective design method of a deep sea water turbine, which comprises the following steps: performing first round optimization and second round optimization on optimization variables of the water turbine, setting a first round optimal solution set as a second round optimization interval, establishing a calculation model between the optimization variables of the water turbine and optimization objectives based on a blade element-momentum theory in the first round optimization, and establishing a second round optimal solution set based on the blade element-momentum theory in the second round optimization interval; establishing a first agent model between the water turbine optimization variable and the optimization target according to a calculation result, and obtaining a first round optimal solution set of the water turbine optimization variable through a multi-target optimization algorithm; in the second round of optimization, a second agent model between the water turbine optimization variable and the optimization target is established on the basis of CFD fluid dynamics calculation or a water tunnel experiment result, and a second round optimal solution set of the water turbine optimization variable is obtained through a multi-target optimization algorithm. And the deep-sea water turbine hydrodynamic configuration with both efficiency and self-starting capability is obtained.

Description

Technical Field

[0001] The invention relates to the field of new energy technology, and in particular to a multi-round multi-objective design method for a deep-sea turbine. Background Art

[0002] Deep sea areas (water depth greater than 200 meters) contain extremely rich mineral resources, which are the key to future human development. Among them, oil and gas resources exceed 38 billion tons, which is four times the proven reserves on land, and the seabed rare earth reserves are 800 times the rare earth reserves on land. Deep-sea observation is the basis for the development of deep-sea resources. However, the power supply problem of deep-sea observation equipment is a bottleneck problem that restricts deep-sea observation. Compared with land-based power supply and ship-based power supply, the use of in-situ energy in the deep sea for distributed power supply of various deep-sea observation devices can greatly reduce the cost of deep-sea monitoring and enable long-term continuous observation of the deep sea in any sea area. Compared with other energy forms in the ocean (salinity difference energy, temperature difference energy, biomass energy, etc.), deep-sea current energy has a higher energy density and a wider distribution range. However, due to the extremely low flow rate in the deep sea and the harsh environment in the deep sea, there is no condition for installing an auxiliary starting device. Therefore, compared with the sea current energy turbine installed in shallow sea areas, the deep-sea sea current energy turbine not only requires a higher conversion efficiency, but also requires good self-starting ability.

[0003] Computational fluid dynamics technology and water tunnel experiments are key technologies for optimizing turbine structures. However, each experiment requires a certain cost and it is difficult to support large-scale optimization processes. Therefore, how to optimize turbine structures quickly and efficiently is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of this, the present invention provides a multi-round multi-objective design method for a deep-sea turbine.

[0005] Specifically, the following technical solutions are included:

[0006] A multi-round multi-objective design method for a deep-sea turbine, comprising:

[0007] Based on environmental parameters and turbine design parameters, determine turbine optimization variables, the first round of optimization interval and optimization target;

[0008] Based on the blade element-momentum theory, a calculation model between the turbine optimization variable and the optimization target is established, and a first proxy model between the turbine optimization variable and the optimization target is established according to the calculation result;

[0009] Obtaining a first round optimal solution set of the turbine optimization variables by a multi-objective optimization algorithm, wherein the first round optimal solution set is a second round optimization interval;

[0010] According to the turbine design parameters, the turbine optimization variables and the second round optimization interval, sampling points are randomly selected based on the optimal Latin hypercube experimental design method to obtain the turbine structure corresponding to each sampling point, and CFD fluid dynamics calculation or water tunnel experiment is performed on the turbine structure, and a second proxy model between the turbine optimization variables and the optimization target is established according to the CFD fluid dynamics calculation or water tunnel experiment results;

[0011] A second round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm.

[0012] Preferably, the turbine optimization variables include: tip speed ratio, maximum relative camber, relative position of maximum relative camber, maximum relative thickness, relative position of maximum relative thickness, chord length, and twist angle;

[0013] The optimization objectives include: energy utilization coefficient and starting torque coefficient.

[0014] Preferably, based on the blade element-momentum theory, a calculation model between the turbine optimization variables and the optimization target is established, including:

[0015] According to the turbine optimization variables and the first round optimization interval, based on a preset airfoil database and an airfoil data calling model, a first lift coefficient database and a first drag coefficient database are obtained;

[0016] Obtaining an estimated angle of attack according to a preset relative flow velocity angle and a preset twist angle;

[0017] According to the estimated angle of attack, searching in the first lift coefficient database and the first drag coefficient database to obtain an estimated first lift coefficient and a first drag coefficient;

[0018] According to the estimated first lift coefficient and the first drag coefficient, an axial induction factor and a circumferential induction factor are calculated according to the blade element-momentum theory;

[0019] A new value of the torsion angle is calculated according to the axial induction factor and the circumferential induction factor, the maximum value between the preset torsion angle and the new value of the torsion angle is selected to update the preset torsion angle, the estimated angle of attack is updated according to the updated preset torsion angle, and a first lift coefficient and a first drag coefficient are determined according to the updated estimated angle of attack;

[0020] The optimization target is obtained according to the determined first lift coefficient and first drag coefficient and by using blade element-momentum theory.

[0021] Preferably, according to the turbine optimization variables and the first round optimization interval, based on a preset airfoil database and an airfoil data calling model, a first lift coefficient database and a first drag coefficient database are obtained, including:

[0022] Obtaining a preset range of angle of attack, and obtaining a first lift coefficient and a first drag coefficient within the preset range of angle of attack based on other turbine optimization variables and the first round of optimization interval and a pre-set airfoil database and airfoil data calling model;

[0023] The first lift coefficient and the first drag coefficient within the preset range of the angle of attack are expanded to the first lift coefficient and the first drag coefficient within the full range of the angle of attack by using the Viterna-Corrigan formula;

[0024] A first lift coefficient database and a first drag coefficient database are obtained according to the first lift coefficient and the first drag coefficient within the full angle of attack range by using a lift and drag coefficient proxy model.

[0025] Preferably, a first proxy model between the turbine optimization variables and the optimization target is established based on the calculation results, including: using Isight software and selecting a radial basis function neural network to establish the first proxy model between the turbine optimization variables and the optimization target.

[0026] Preferably, the first round optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm, including: optimizing the first proxy model in the first round optimization space based on the NSGA-Ⅱ multi-objective optimization algorithm to obtain the Pareto optimal solution set of the turbine optimization variables.

[0027] Preferably, sampling points are randomly selected based on the optimal Latin hypercube experimental design method according to the turbine design parameters, turbine optimization variables and the second round optimization interval, including: obtaining an experimental design table according to the sampling points.

[0028] Preferably, obtaining a turbine structure corresponding to each sampling point and performing CFD fluid dynamics calculation or water tunnel experiment on the turbine structure includes:

[0029] A turbine model is established according to the turbine structure, the turbine model is meshed and discretized, a velocity inlet-pressure outlet boundary condition is adopted, and the torque and speed of the turbine model are monitored to obtain a CFD fluid dynamics calculation result of the optimization target;

[0030] Fill in the experimental design table with the CFD fluid dynamics calculation results of the optimization target corresponding to each sampling point.

[0031] Preferably, a second proxy model between the turbine optimization variables and the optimization target is established based on CFD fluid dynamics calculations or water tunnel experiment results, including: using Isight software and selecting a radial basis function neural network to establish a second proxy model between the turbine optimization variables and the optimization target.

[0032] Preferably, the second round optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm, including: optimizing the second proxy model in the second round optimization space based on the NSGA-Ⅱ multi-objective optimization algorithm to obtain the Pareto optimal solution set of the turbine optimization variables.

[0033] The beneficial effects of the technical solution provided by the present invention include at least:

[0034] The present application performs a first round of optimization and a second round of optimization on the turbine optimization variables, and the first round of optimal solution set is the second round of optimization interval. The first round of optimization is based on the blade element-momentum theory, and a calculation model between the turbine optimization variables and the optimization target is established. According to the calculation results, a first proxy model between the turbine optimization variables and the optimization target is established, and the first round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm. The second round of optimization is based on CFD fluid dynamics calculations or water tunnel experiment results to establish a second proxy model between the turbine optimization variables and the optimization target, and the second round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm. The present application can perform large-scale optimization at a very low cost to obtain a deep-sea turbine hydrodynamic configuration with both efficiency and self-starting capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0036] Figure 1 is a schematic flow chart of a multi-round multi-objective design method for a deep-sea turbine in one embodiment of the present invention;

[0037] Figure 2 is a schematic diagram of the structure of a water turbine in one embodiment of the present invention;

[0038] Figure 3 1 is a schematic diagram of the range of turbine optimization variables in the first round of optimal solution set in one embodiment of the present invention;

[0039] Figure 4 It is a schematic diagram of a structured grid generated by a turbine model in a CFD fluid dynamics calculation in one embodiment of the present invention;

[0040] Figure 5 is a schematic diagram of the flow field structure of a numerical model of a water turbine model in one embodiment of the present invention;

[0041] Figure 6This is a schematic diagram of the structure of a water tunnel test bench in one embodiment of the present invention;

[0042] Figure 7 It is a schematic diagram of the Pareto optimal solution set of the turbine optimization variables obtained in the second round in step S50 in one embodiment of the present invention.

[0043] The above drawings have shown clear embodiments of the present invention, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present invention in any way, but to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] As mentioned above, existing fluid mechanics technology and water tunnel experiments are key technologies for optimizing turbine structures. However, each experiment requires a certain cost and is difficult to support large-scale optimization processes. In order to solve the above technical problems, an embodiment of the present invention provides a multi-round multi-objective design method for deep-sea turbines, such as Figure 1 As shown, including:

[0046] S10: Based on environmental parameters and turbine design parameters, determine turbine optimization variables, first round optimization interval and optimization target;

[0047] S20: Based on the blade element-momentum theory, a calculation model between the turbine optimization variable and the optimization target is established, and a first proxy model between the turbine optimization variable and the optimization target is established according to the calculation result;

[0048] S30: obtaining a first round optimal solution set of the turbine optimization variables by a multi-objective optimization algorithm, wherein the first round optimal solution set is a second round optimization interval;

[0049] S40: According to the turbine design parameters, the turbine optimization variables and the second round optimization interval, sampling points are randomly selected based on the optimal Latin hypercube test design method to obtain the turbine structure corresponding to each sampling point, and CFD fluid dynamics calculation or water tunnel experiment is performed on the turbine structure, and a second proxy model between the turbine optimization variables and the optimization target is established according to the CFD fluid dynamics calculation or water tunnel experiment results;

[0050] S50: Obtaining a second round of optimal solution set of the turbine optimization variables through a multi-objective optimization algorithm.

[0051] The present application performs a first round of optimization and a second round of optimization on the turbine optimization variables, and the first round of optimal solution set is the second round of optimization interval. The first round of optimization is based on the blade element-momentum theory, and a calculation model between the turbine optimization variables and the optimization target is established. According to the calculation results, a first proxy model between the turbine optimization variables and the optimization target is established, and the first round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm. The second round of optimization is based on CFD fluid dynamics calculations or water tunnel experiment results to establish a second proxy model between the turbine optimization variables and the optimization target, and the second round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm. The present application can perform large-scale optimization at a very low cost to obtain a deep-sea turbine hydrodynamic configuration with both efficiency and self-starting capability.

[0052] Specifically, the environmental parameters in step S10 include: design incoming flow velocity, Reynolds number. The turbine design parameters include: impeller radius, number of blades, hub radius, distance between blade element and center of circle. The turbine optimization variables include: tip speed ratio, maximum relative curvature, relative position of maximum relative curvature, maximum relative thickness, relative position of maximum relative thickness, chord length, torsion angle; the optimization targets include: energy utilization coefficient and starting torque coefficient.

[0053] Specifically, in one embodiment, in step S10, the tip speed ratio, maximum relative camber, relative position of maximum relative camber, maximum relative thickness, relative position of maximum relative thickness, chord length, and twist angle can be set to preset value ranges respectively.

[0054] Specifically, in one embodiment, the turbine optimization variables in step S10 include: the formula corresponding to the maximum relative camber and its relative position changing with the span direction, the formula corresponding to the maximum relative thickness and its relative position changing with the span direction, the formula corresponding to the chord length changing with the span direction, and the formula corresponding to the twist angle changing with the span direction. Considering the superior properties of the polynomial coefficients, the above formulas are all set as third-order polynomials, and the preset ranges of the maximum relative camber, the relative position of the maximum relative camber, the maximum relative thickness, the relative position of the maximum relative thickness, the chord length, and the twist angle are obtained by setting the range of the polynomial coefficients in each polynomial. The span direction is the length extension direction of the blade, and the independent variable in the third-order polynomial is the distance between the blade element and the center of the circle.

[0055] Since the blade element-momentum theory calculates a single blade element, the blade can be divided into multiple blade elements according to the needs, and each blade element is optimized in the first round by the blade element-momentum theory. The number of blade elements can be set according to the needs, which can be thousands or tens of thousands. The turbine optimization variables in each blade element are set separately. It should be noted that in each optimization process of each blade element, the optimization is carried out within the full range of the tip speed ratio, and the tip speed ratio when the energy utilization coefficient takes the maximum value is selected.

[0056] Specifically, in step S20, based on the blade element-momentum theory, a calculation model between the turbine optimization variable and the optimization target is established, including:

[0057] S21: obtaining a first lift coefficient database and a first drag coefficient database according to the turbine optimization variables and the first round optimization interval based on a preset airfoil database and an airfoil data calling model;

[0058] Specifically, for example, the airfoil data calling model is Xfoil software, the airfoil database is NACA airfoil database, and the first lift coefficient database and the first drag coefficient database are obtained in the NACA airfoil database according to the turbine optimization variables and the first round optimization interval using Xfoil software.

[0059] Specifically, they include:

[0060] S211: Obtaining a preset range of angle of attack, and obtaining a first lift coefficient and a first drag coefficient within the preset range of angle of attack based on a preset airfoil database and airfoil data calling model according to other turbine optimization variables and the first round optimization interval;

[0061] Specifically, the preset range of the angle of attack is determined according to the situation of the airfoil database, and is generally a small angle of attack range (-15°, 15°). For example, in the NACA data parameters, the preset range of the maximum relative camber is set to 0-5, the preset range of the relative position of the maximum relative camber is set to 0-3, the preset range of the maximum relative thickness is set to 0-25, the Reynolds number is 20000-900000, and the preset range of the angle of attack is -15°-15°. Using the Xfoil software in the NACA airfoil database, 20 groups of hydrodynamic data of NACA airfoils in the small angle of attack range are obtained, and the first lift coefficient and the first drag coefficient in the small angle of attack range can be obtained according to the hydrodynamic data.

[0062] S212: using the Viterna-Corrigan formula to expand the first lift coefficient and the first drag coefficient within the preset range of the angle of attack to the first lift coefficient and the first drag coefficient within the full range of the angle of attack;

[0063] Specifically, the full angle of attack range is -15°-90°, and the Viterna-Corrigan formula is used to expand the hydrodynamic data in the small angle of attack range to the hydrodynamic data in the full angle of attack range, and the first lift coefficient and the first drag coefficient in the full angle of attack range are obtained according to the hydrodynamic data in the full angle of attack range.

[0064] Both the lift coefficient and the drag coefficient are related to the Reynolds number and the turbine optimization variables.

[0065] S213: Obtaining a first lift coefficient database and a first drag coefficient database according to the first lift coefficient and the first drag coefficient within the full angle of attack range by using a lift and drag coefficient proxy model.

[0066] Specifically, the lift-drag coefficient proxy model is a proxy model that converts a discrete lift coefficient into a continuous lift coefficient and a discrete drag coefficient into a continuous drag coefficient using a fitting method. It can be a fitting formula or a neural network model trained based on historical lift coefficients and drag coefficients.

[0067] S22: Obtaining an estimated angle of attack according to a preset relative flow velocity angle and a preset twist angle;

[0068] Specifically, it is obtained by the following formula:

[0069]

[0070] S23: searching the first lift coefficient database and the first drag coefficient database to obtain an estimated first lift coefficient and a first drag coefficient according to the estimated angle of attack;

[0071] Specifically, the estimated first lift coefficient is obtained in the first lift coefficient database according to the estimated angle of attack, other turbine optimization variables, and the first round of optimization interval, and the estimated first drag coefficient is obtained in the first drag coefficient database.

[0072] S24: according to the estimated first lift coefficient and the first drag coefficient, calculate and obtain an axial induction factor and a circumferential induction factor according to the blade element-momentum theory;

[0073] Specifically, according to momentum theory, the thrust element and the thrust torque element are obtained from the axial momentum and the angular momentum, respectively, as shown in Equation 2 and Equation 3.

[0074] dT=4a(1-a)ρU 2 πr dr (2)

[0075] dQ=4a′(1-a)ρUπr 3 Ωdr (3)

[0076] According to the blade element theory, the thrust element and thrust moment element are obtained as shown in Equation 4 and Equation 5 respectively.

[0077]

[0078] According to the blade element-momentum theory, the axial induction factor a and the circumferential induction factor a are obtained. ′ As shown in Equation 6 and Equation 7.

[0079]

[0080] Where: F is the loss correction factor, as shown in Formula 8.

[0081]

[0082] S24: obtaining a new value of the torsion angle by calculating according to the axial induction factor and the circumferential induction factor, selecting the maximum value between the preset torsion angle and the new value of the torsion angle to update the preset torsion angle, updating the estimated angle of attack according to the updated preset torsion angle, and obtaining a determined first lift coefficient and a first drag coefficient according to the updated estimated angle of attack;

[0083] Specifically, the new value of the torsion angle Calculated by formula 9:

[0084]

[0085] Compare and The size of Compare If big, use Steps S22 to S23 are recalculated to obtain an updated preset twist angle to update the estimated angle of attack, and the first lift coefficient and the first drag coefficient are determined according to the updated estimated angle of attack. Compare If the value is small, directly execute the next step S25.

[0086] S25: According to the determined first lift coefficient and first drag coefficient, the optimization target is obtained by blade element-momentum theory calculation.

[0087] Specifically, according to the supplementary equations 10, 11 and 12, the optimization objectives are obtained respectively: the first energy utilization coefficient C p and the first starting torque coefficient C t,s The expressions of are shown in Equation 13 and Equation 14.

[0088]

[0089] In the above formulas 1-14, the meanings of the parameters are shown in Table 1. The turbine structure is as follows: Figure 2 As shown:

[0090] Table 1 Parameter selection of NSGA-Ⅱ multi-objective optimization algorithm

[0091]

[0092]

[0093] Specifically, in step S20, a first proxy model between the turbine optimization variables and the optimization target is established according to the calculation results, including: using Isight software and selecting a radial basis function neural network to establish a first proxy model between the turbine optimization variables and the optimization target.

[0094] Specifically, using Isight software, selecting the radial basis function neural network, adjusting the smoothness to 0, the first proxy model was obtained. The first proxy model had excellent results, with average errors of 0.0008 and 0.0002, respectively, and the variance R 2 The input of the radial basis function neural network is the turbine optimization variable, and the output of the radial basis function neural network is the optimization target.

[0095] Specifically, in step S30, a first-round optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm, including: optimizing the first proxy model in the first-round optimization space based on the NSGA-Ⅱ multi-objective optimization algorithm to obtain the Pareto optimal solution set of the turbine optimization variables.

[0096] Specifically, the parameter selection of NSGA-Ⅱ multi-objective optimization algorithm is shown in Table 2. The optimization goal is the maximum energy utilization coefficient C p,max and starting torque coefficient C t,s . C p,max and C t,s The goal is to maximize the energy utilization coefficient at all tip speed ratios and the starting torque coefficient when the tip speed ratio is 0.

[0097] Table 2 Parameter selection of NSGA-Ⅱ multi-objective optimization algorithm

[0098]

[0099] Specifically, as a specific embodiment, the input data - turbine optimization variables, first round optimization interval are obtained according to the user's instructions, and the first lift coefficient database and the first drag coefficient database are obtained according to step S21; the distance r between the blade element and the center of the circle and the preset relative flow velocity angle are obtained according to the user's instructions. The distance between the blade element and the center of the circle is R hub , R hub The radius at the blade root is 0.1m, and the relative velocity angle is preset is 10°, and the estimated angle of attack is obtained according to formula 1; and steps S22-S25 are executed to obtain the first energy utilization coefficient C of the first round of optimization target. p and the first starting torque coefficient C t,s Through the multi-objective optimization algorithm of step S30, a total of 2881 groups of working conditions are calculated to obtain the Pareto optimal solution set of turbine optimization variables, which is a total of 41 groups. The range of turbine optimization variables in the first round of optimal solution set is as follows: Figure 3 As shown, this range is used as the second round of optimization interval. Figure 3 Where a01, b01, and c01 are the polynomial coefficients of the third-order polynomial corresponding to the curve of the change of chord length with the span direction, a02, b02, and c02 are the polynomial coefficients of the third-order polynomial corresponding to the curve of the change of torsion angle with the span direction, houdu is the maximum relative thickness, lamda is the tip speed ratio, wandu is the maximum relative camber, and wandusit is the relative position of the maximum relative camber.

[0100] Specifically, in step S40, sampling points are randomly selected based on the optimal Latin hypercube experimental design method according to the turbine design parameters, turbine optimization variables and the second round optimization interval, including: S41: obtaining an experimental design table according to the sampling points. This setting can ensure that all the sampled sampling points can be evenly distributed in the second round optimization interval. Specifically, in this embodiment, 300 sampling points are selected in the second round optimization interval, and the obtained experimental design table is shown in Table 3.

[0101] Table 3 Experimental design table

[0102] 1 2 3 · · · · · · · · · · · 300 <![CDATA[c1]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[β1]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdu1]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdusit1]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandu1]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandusit1]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[c2]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[β2]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdu2]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdusit2]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandu2]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandusit2]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[c3]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[β3]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdu3]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdusit3]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandu3]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandusit3]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[c4]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[β4]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdu4]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Houdusit4]]> xx xx xx xx xx xx xx xx xx xx xx xx xx xx xx <![CDATA[Wandu4]]> xx xx xx xx xx xx xx [[ID=348 ​ ​ ​ ​ ​ ​ ​ <![CDATA[Wandusit4]]> ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ <![CDATA[C p,max ]]> <![CDATA[C t,s ]]>

[0103] It should be noted that in Table 3, c represents the chord length, β represents the torsion angle, Houdu represents the maximum relative thickness, Houdusit represents the relative position of the maximum relative thickness, Wandu represents the maximum relative curvature, Wandusit represents the relative position of the maximum relative curvature, subscript 1 is the parameter at a distance of 0.25 between the blade element and the center of the circle, subscript 2 is the parameter at a distance of 0.5 between the blade element and the center of the circle, subscript 3 is the parameter at a distance of 0.75 between the blade element and the center of the circle, subscript 4 is the parameter at a distance of 1 between the blade element and the center of the circle, and xx represents a specific value. A column of data in Table 3 is data for a sampling point. Table 3 is for illustration only and does not impose any restrictions on this application.

[0104] Specifically, the distance between the blade element and the center of the circle in the experimental design table can be set according to needs.

[0105] Specifically, in step S40, a turbine structure corresponding to each sampling point is obtained, and a CFD fluid dynamics calculation or a water tunnel experiment is performed on the turbine structure, including:

[0106] S42: establishing a turbine model according to the turbine structure, performing mesh discretization on the turbine model, adopting a velocity inlet-pressure outlet boundary condition, and monitoring the torque and speed of the turbine model to obtain a CFD fluid dynamics calculation result of the optimization target;

[0107] Specifically, the turbine structure is obtained according to the data and lofting in Table 3. The turbine model is established using the geometric modeling software Spaceclaim, and the structured grid of the turbine model is generated using the NUMECA software, such as ​ As shown, the number of generated meshes is about 2 million. The mesh file is imported into ANSYS Fluent 2020, the velocity inlet-pressure outlet boundary setting is adopted, and the torque and speed of the turbine are monitored to obtain the second maximum energy utilization coefficient C p,max and the second starting torque coefficient C t,s The flow field structure of the numerical model of the turbine model is as follows ​ shown.

[0108] Alternatively, a physical model of the turbine structure corresponding to each sampling point is made and the model is constructed on a water tunnel test bench (such as ​ The experiment is shown in Figure 2 to obtain the hydrodynamic performance data of the turbine, and then the second maximum energy utilization coefficient C is obtained. p,max and the second starting torque coefficient C t,s data.

[0109] S43: Fill in the CFD fluid dynamics calculation results of the optimization target corresponding to each sampling point in the experimental design table.

[0110] Specifically, the optimization objectives are filled in the last two columns in Table 3.

[0111] Specifically, as shown in Table 3, the sampling points in Table 3 are discrete points. For subsequent optimization, in step S40, a second proxy model between the turbine optimization variables and the optimization target is established according to the CFD fluid dynamics calculation or water tunnel test results, including: using Isight software, selecting a radial basis function neural network to establish a second proxy model between the turbine optimization variables and the optimization target. Specifically, using Isight software, selecting a radial basis function neural network, adjusting the smoothness to 0, and obtaining a second proxy model. The second proxy model has excellent results, with average errors of 0.0008 and 0.0002, respectively, and the variance R 2 The input of the radial basis function neural network is the turbine optimization variable, and the output of the radial basis function neural network is the optimization target.

[0112] Step S50 obtains the second round optimal solution set of the turbine optimization variables through a multi-objective optimization algorithm, including: optimizing the second proxy model in the second round optimization space based on the NSGA-Ⅱ multi-objective optimization algorithm to obtain the Pareto optimal solution set of the turbine optimization variables. Specifically, the parameter selection of the NSGA-Ⅱ multi-objective optimization algorithm is shown in Table 2.

[0113] The Pareto optimal solution set obtained in step S50 is as follows: ​ As shown, ​ Taking the optimum point 2 in the example as an example, the large turbine can obtain a starting torque coefficient C of 0.09 by using the present application. t,s At the same time, the maximum energy utilization coefficient C of more than 0.32 is obtained p,max In order to verify the effectiveness of the design method in the present invention, this application also passes ​ The water tunnel test bench shown is used to verify the performance of the turbine, and the obtained experimental data proves the effectiveness of the optimization method described in the present invention.

[0114] ​ The water tunnel test bench includes a circulating flow water tank, an experimental turbine, a bracket, a slider, a slide rail, and an experimental platform. The circulating flow water tank is filled with water, the experimental turbine is immersed in the circulating flow water tank, the experimental turbine is installed at the lower end of the experimental platform, the lower end of the experimental platform is immersed in the circulating flow water tank, the upper end of the experimental platform is located on the upper side of the top cover of the circulating flow water tank, the top cover of the circulating flow water tank is provided with a slide rail, the upper part of the experimental platform is connected to the slider, and the slider is slidably connected to the slide rail. The cooperation of the slide rail and the slider can realize the horizontal movement of the experimental platform.

[0115] The present application performs a first round of optimization and a second round of optimization on the turbine optimization variables, and the first round of optimal solution set is the second round of optimization interval. The first round of optimization is based on the blade element-momentum theory, and a calculation model between the turbine optimization variables and the optimization target is established. According to the calculation results, a first proxy model between the turbine optimization variables and the optimization target is established, and the first round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm. The second round of optimization is based on CFD fluid dynamics calculations or water tunnel experiment results to establish a second proxy model between the turbine optimization variables and the optimization target, and the second round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm. The present application can perform large-scale optimization at a very low cost to obtain a deep-sea turbine hydrodynamic configuration with both efficiency and self-starting capability.

[0116] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A multi-round multi-objective design method for a deep-sea turbine, characterized in that: include: Based on environmental parameters and turbine design parameters, determine turbine optimization variables, the first round of optimization interval and optimization target; Based on the blade element-momentum theory, a calculation model between the turbine optimization variable and the optimization target is established, and a first proxy model between the turbine optimization variable and the optimization target is established according to the calculation result; Obtaining a first round optimal solution set of the turbine optimization variables by a multi-objective optimization algorithm, wherein the first round optimal solution set is a second round optimization interval; According to the turbine design parameters, the turbine optimization variables and the second round optimization interval, sampling points are randomly selected based on the optimal Latin hypercube experimental design method to obtain the turbine structure corresponding to each sampling point, and CFD fluid dynamics calculation or water tunnel experiment is performed on the turbine structure, and a second proxy model between the turbine optimization variables and the optimization target is established according to the CFD fluid dynamics calculation or water tunnel experiment results; A second round of optimal solution set of the turbine optimization variables is obtained through a multi-objective optimization algorithm.

2. A multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: The turbine optimization variables include: tip speed ratio, maximum relative camber, relative position of maximum relative camber, maximum relative thickness, relative position of maximum relative thickness, chord length, and twist angle; The optimization objectives include: energy utilization coefficient and starting torque coefficient.

3. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: Based on the blade element-momentum theory, a calculation model between the turbine optimization variables and the optimization target is established, including: According to the turbine optimization variables and the first round optimization interval, based on a preset airfoil database and an airfoil data calling model, a first lift coefficient database and a first drag coefficient database are obtained; Obtaining an estimated angle of attack according to a preset relative flow velocity angle and a preset twist angle; According to the estimated angle of attack, searching in the first lift coefficient database and the first drag coefficient database to obtain an estimated first lift coefficient and a first drag coefficient; According to the estimated first lift coefficient and the first drag coefficient, an axial induction factor and a circumferential induction factor are calculated according to the blade element-momentum theory; A new value of the torsion angle is calculated according to the axial induction factor and the circumferential induction factor, the maximum value between the preset torsion angle and the new value of the torsion angle is selected to update the preset torsion angle, the estimated angle of attack is updated according to the updated preset torsion angle, and a first lift coefficient and a first drag coefficient are determined according to the updated estimated angle of attack; The optimization target is obtained according to the determined first lift coefficient and first drag coefficient and by using blade element-momentum theory.

4. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: According to the turbine optimization variables and the first round optimization interval, based on the preset airfoil database and the airfoil data calling model, a first lift coefficient database and a first drag coefficient database are obtained, including: Obtaining a preset range of angle of attack, and obtaining a first lift coefficient and a first drag coefficient within the preset range of angle of attack based on other turbine optimization variables and the first round of optimization interval and a pre-set airfoil database and airfoil data calling model; The first lift coefficient and the first drag coefficient within the preset range of the angle of attack are expanded to the first lift coefficient and the first drag coefficient within the full range of the angle of attack by using the Viterna-Corrigan formula; A first lift coefficient database and a first drag coefficient database are obtained according to the first lift coefficient and the first drag coefficient within the full angle of attack range by using a lift and drag coefficient proxy model.

5. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: A first proxy model between the turbine optimization variables and the optimization target is established according to the calculation results, including: using Isight software and selecting a radial basis function neural network to establish the first proxy model between the turbine optimization variables and the optimization target.

6. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: The first round optimal solution set of the turbine optimization variables is obtained by a multi-objective optimization algorithm, including: optimizing the first proxy model in the first round optimization space based on the NSGA-Ⅱ multi-objective optimization algorithm to obtain the Pareto optimal solution set of the turbine optimization variables.

7. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: According to the turbine design parameters, turbine optimization variables and the second round optimization interval, sampling points are randomly selected based on the optimal Latin hypercube experimental design method, including: obtaining an experimental design table according to the sampling points.

8. A multi-round multi-objective design method for a deep-sea turbine according to claim 7, characterized in that: The turbine structure corresponding to each sampling point is obtained, and CFD fluid dynamics calculation or water tunnel experiment is performed on the turbine structure, including: A turbine model is established according to the turbine structure, the turbine model is meshed and discretized, a velocity inlet-pressure outlet boundary condition is adopted, and the torque and speed of the turbine model are monitored to obtain a CFD fluid dynamics calculation result of the optimization target; Fill in the experimental design table with the CFD fluid dynamics calculation results of the optimization target corresponding to each sampling point.

9. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: A second proxy model between the turbine optimization variables and the optimization target is established according to CFD fluid dynamics calculations or water tunnel test results, including: using Isight software and selecting a radial basis function neural network to establish a second proxy model between the turbine optimization variables and the optimization target.

10. The multi-round multi-objective design method for a deep-sea turbine according to claim 1, characterized in that: The second round optimal solution set of the turbine optimization variables is obtained by a multi-objective optimization algorithm, including: optimizing the second proxy model in the second round optimization space based on the NSGA-Ⅱ multi-objective optimization algorithm to obtain the Pareto optimal solution set of the turbine optimization variables.

Images