Aerodynamic design method for biomimetic tower-type wind turbine
By using a biomimetic tower design, combined with numerical simulation and experimental measurement, the seal whisker model was improved into a wind turbine tower, solving the problems of vortex-induced vibration and tower shadow effect, and improving the power output and yaw function of the wind turbine.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- YANGZHOU UNIV
- Filing Date
- 2023-03-15
- Publication Date
- 2026-05-05
AI Technical Summary
The tower of modern large wind turbines is prone to vortex-induced vibration and tower shadow effect in wind flow, which affects the fatigue life of the blades and the output power. In addition, the wake region of the wind turbine affects the power output of downstream wind turbines.
By adopting a biomimetic tower design and combining numerical simulation and experimental measurement, the seal whisker model was modified into a wind turbine tower. Through 3D modeling, 3D printing and wind tunnel testing, the aerodynamic characteristics of different shapes and structures were studied to suppress vortex shedding and tower shadow effect and optimize the aerodynamic performance of the wind turbine.
It effectively suppresses vortex-induced vibration, reduces the aerodynamic load of the tower on the blades, improves the power output and yaw function of the wind turbine, and enhances the overall performance of the wind turbine.
Smart Images

Figure CN116263837B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wind turbine aerodynamic characteristic optimization design, and particularly to an aerodynamic design method for a biomimetic tower-type wind turbine. Background Technology
[0002] Energy is inextricably linked to human society, serving as a crucial material foundation for developing production and improving living standards. With population growth and economic development, the demand for energy is constantly increasing. In today's world, with fossil fuel shortages and increasingly severe global climate change, developing clean and renewable energy has become one of the most pressing energy issues internationally. Wind energy, as a sustainable and renewable energy source, is clean, pollution-free, and possesses significant ecological and environmental benefits. The global exploitable wind resources amount to approximately 4350 GW, of which about 1000 GW is onshore and 200 GW is offshore, available for commercial development using existing technologies. China has relatively abundant resources, with its vast territory and long, winding coastline providing rich wind resources. Significantly increasing the proportion of wind energy utilization in my country is a crucial measure for achieving the "3060" national energy strategy.
[0003] With the rapid development of the global wind turbine industry, the trend towards larger wind turbines has become increasingly prominent. Modern large-scale megawatt-class horizontal axis variable pitch constant frequency wind turbines mostly employ a cylindrical thin-walled tower design. However, in practical engineering applications, wind is highly variable, exhibiting significant spatial and temporal uncertainty. This makes the impact of wind on the turbine blades and tower structure, as well as the loads they experience, extremely complex. When fluid passes through the wind turbine, frequent detachment of Karman vortex streets occurs behind the tower. The interaction between these vortex streets and the tower body generates vortex-induced vibrations, a phenomenon that poses serious safety hazards to the tower structure. Furthermore, the tower shadow effect, caused by the tower's obstruction, affects the aerodynamic loads on the blades, reducing their fatigue life and output power. The tower shadow effect also influences the wind turbine's wake region, further impacting the power output of downstream wind turbines.
[0004] Bionic research indicates that the flow field around a seal's whiskers, compared to a smooth cylinder, exhibits significantly reduced average and pulsating drag, markedly weakened whisker vibration, and substantial suppression of vortex shedding, which helps reduce pulsating pressure on the cylinder surface. Therefore, an aerodynamic design for a biomimetic tower-type wind turbine is proposed. Using a combination of experimental measurements and numerical calculations, the influence mechanism of the biomimetic tower on the load reduction and power increase of horizontal axis wind turbine blades is systematically studied. This approach modifies the drag characteristics and wake structure of traditional cylindrical towers, reducing the tower shadow effect on the blades to ensure the wind turbine can minimize safety hazards and maximize power output. Summary of the Invention
[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an aerodynamic design method for biomimetic tower-type wind turbines. This method is applicable to the aerodynamic design of wind turbine towers under different operating conditions, improves the influence of the tower shadow effect, increases the power output of wind turbines, provides a theoretical basis and scientific basis for the aerodynamic optimization design of horizontal axis wind turbines, and is beneficial to the practical application of wind turbines.
[0006] The objective of this invention is achieved as follows: an aerodynamic design method for a biomimetic tower-type wind turbine, comprising the following steps:
[0007] Step 1) Improve the design of the wind turbine tower using 3D modeling software. Improve the existing seal whisker model and select a traditional smooth cylindrical tower with the same hydraulic diameter for comparative study.
[0008] Step 2) Using numerical simulation, the improved biomimetic tower model is compared with the cylindrical tower model. The aerodynamic characteristics of towers with different shapes and structures are studied from the aspects of surface force, velocity distribution, flow mode and pressure distribution.
[0009] Step 3) 3D print the improved biomimetic tower column model and cylindrical tower column model, and evenly distribute pressure measuring holes at the nodal section and saddle point section;
[0010] Step 4) Set up the wind tunnel test setup and conduct the test using an electronic pressure scanning valve and a hot-wire anemometer;
[0011] Step 5) Use wind tunnel testing to collect data, compare the surface pressure, aerodynamic forces and wakes of the cylindrical tower model and the biomimetic tower model, and study the aerodynamic characteristics of towers with different shapes and structures.
[0012] Step 6) Use 3D modeling software to model the wind turbine blades, hub, and nacelle, and combine the designed biomimetic tower with the rotating blades to apply it to a horizontal axis wind turbine.
[0013] Step 7) Using numerical simulation, a series of simulation results are compared between the aerodynamic performance and wake field characteristics of the biomimetic tower-type wind turbine and the cylindrical tower-type wind turbine. The study investigates the influence of the tower shadow effect on the load on the blades, the axial force on the rotating wind turbine, and the instantaneous torque for towers with different shapes and structures.
[0014] As a further limitation of the present invention, the improved design of the wind turbine tower described in step 1) specifically includes: changing the major and minor axis ratios of the two elliptical control sections and the tilt angle with the horizontal plane; taking the elliptical control section at the center height of the tower as the base plane, the elliptical control sections above the base plane are rotated counterclockwise by an angle γ relative to the next adjacent control plane around the center line of the tower height direction; the elliptical control sections below the base plane are rotated clockwise by an angle γ relative to the previous adjacent control plane around the center line of the tower height direction, so that the tower is spiral-shaped along the height direction.
[0015] As a further limitation of the present invention, the numerical simulation calculation method described in steps 2) and 7) specifically includes the calculation condition being Reynolds number R. e In the subcritical Reynolds number region of 30000, the applied turbulence model is based on γ-Reynolds number. θ A four-equation transition model is formed by coupling the transition model and the k-ωSST model. The simulation conditions are different wind attack angles. The biomimetic tower model and the cylindrical tower model are compared and the calculation and analysis are carried out from the aspects of surface force, velocity distribution, flow mode and pressure distribution of the tower.
[0016] As a further limitation of the present invention, the method of using wind tunnel testing in step 5) specifically includes, under the calculated operating condition of Reynolds number R... e Data is collected in the subcritical Reynolds number region of 30,000. The bottom of the wind tunnel test device has a pre-set opening to facilitate the connection of the pressure hose to the electronic pressure acquisition system. A disc is installed on the top of the test column to eliminate interference at the end.
[0017] As a further limitation of the present invention, the surface pressure and aerodynamic force measurement of the cylindrical tower model in step 5) is processed using formula (1);
[0018]
[0019] In equation (1), p j The static pressure at test port j is j = 1~19, p0 is the incoming static pressure measured by the Pitot tube, ρ is the incoming air density, U0 is the free flow velocity, and C... p This is the pressure coefficient of the measuring hole on the cross section.
[0020] As a further limitation of the present invention, the surface pressure and aerodynamic force measurement of the biomimetic tower model in step 5) is processed using formulas (2) and (3);
[0021]
[0022]
[0023] In equations (2) and (3), Cl Let C be the lift coefficient of each section of the biomimetic cylinder. d θ represents the drag coefficient of each section of the biomimetic column, θ is the angle between the measuring point and the original stationary point, and its value ranges from 0° to 180°. N represents the measuring points arranged within the range of 0° to 180°.
[0024] Compared with existing technologies, the beneficial effects of this invention, which adopts the above technical solutions, are as follows: For the aerodynamic design of wind turbine towers, this invention improves upon the existing seal whisker model, using a combination of numerical simulation and experimental measurement to verify the correctness of the design concept. The experiments demonstrate good stability and high data accuracy. Furthermore, the aerodynamic forces and vortex-induced vibration suppression capabilities of the biomimetic column with a similar major and minor axes are analyzed. Subsequently, this invention employs numerical simulation to combine the designed biomimetic tower with rotating blades and applies it to a horizontal axis wind turbine. Through comparative analysis of numerical simulation calculations of the biomimetic tower and cylindrical tower under blade disturbance, the aerodynamic load on the blades caused by the tower shadow effect due to the different tower shapes at different blade azimuth angles is studied. The influence of the biomimetic tower on the axial force and instantaneous torque of the rotating wind turbine is analyzed. The main consideration in this invention is that, in practical applications, when wind blows over the tower, vortex streets alternately break off on both sides of the wake region behind the tower. The interaction between the vortex streets and the tower can lead to vortex-induced vibration, a phenomenon that poses a serious safety hazard to the tower's shape. Therefore, this invention uses a modified seal whisker-inspired bionic tower as the wind turbine tower, effectively suppressing vortex street breakoff behind the tower. Considering the tower's sensitivity to incoming wind from all directions, and the significant variations in lift and drag on the tower surface due to different wind directions, resulting in poor yaw performance, this invention uses a modified bionic tower for numerical simulation and wind tunnel testing to verify that the designed bionic tower can effectively apply yaw functionality under different wind angles of attack. Considering that the rotational disturbance of wind turbine blades will affect the wake field behind the tower, and that the tower shadow effect caused by the tower will affect the aerodynamic load, thereby affecting the thrust and torque on the blades and the output power of the wind turbine, a biomimetic tower-type wind turbine aerodynamic design is adopted to reduce the impact of the tower shadow effect on the aerodynamic load and the tower wake region, thereby improving the power output of the wind turbine. Attached Figure Description
[0025] Figure 1 This is a schematic diagram of the ideal seal whisker columnar structure in this invention.
[0026] Figure 2 This is a structural diagram of the improved biomimetic tower model in this invention.
[0027] Figure 3 (a) is a comparison of the surface resistance coefficients of the biomimetic tower column and the cylindrical tower column in this invention; Figure 3(b) is a comparison of the surface lift coefficients of the biomimetic tower column and the cylindrical tower column in this invention.
[0028] Figure 4 This is a schematic diagram of wind tunnel test measurement in this invention.
[0029] Figure 5 (a) is a comparison of the time-averaged pressure on the surface of the wind tunnel test tower column in this invention; Figure 5 (b) is a comparison of the standard deviation of the surface pressure of the wind tunnel test tower column in this invention.
[0030] Figure 6 This is a diagram showing the aerodynamic verification test results of the biomimetic tower column in the wind tunnel test of this invention.
[0031] Figure 7 This is a structural diagram of the biomimetic tower-type wind turbine model designed in this invention.
[0032] Figure 8 (a) is a graph showing the torque variation with azimuth angle for the biomimetic tower-type wind turbine and the cylindrical tower-type wind turbine in this invention; Figure 8 (b) is a graph showing the thrust of the biomimetic tower-type wind turbine and the cylindrical tower-type wind turbine in this invention as a function of azimuth angle. Detailed Implementation
[0033] An aerodynamic design method for a biomimetic tower-type wind turbine includes the following steps:
[0034] Step 1) Improve the design of the wind turbine tower using 3D modeling software. Improve the existing seal whisker model and select a traditional smooth cylindrical tower with the same hydraulic diameter for comparative study.
[0035] The ideal seal whisker columnar structure for specific modeling can be as follows: Figure 1 As shown. Figure 1 The image shows the horizontal cross-sectional shape of the column and its overall shape with a periodic wave-like appearance; the horizontal cross-section with the larger ratio of the major and minor axes is defined as the nodal cross-section, i.e., the N plane; the horizontal cross-section with the smaller ratio of the major and minor axes is defined as the saddle point cross-section, i.e., the S plane; this ideal seal whisker model is named Whiskers #1.
[0036] The specific design improvements to the biomimetic tower model structure can be as follows: Figure 2As shown. By changing the major and minor axis ratios and the tilt angle with the horizontal plane of the two elliptical control sections, and taking the elliptical control section at the center height of the tower as the base plane, the elliptical control sections above the base plane are rotated counterclockwise by an angle γ relative to the next adjacent control plane about the center line of the tower height, while the elliptical control sections below the base plane are rotated clockwise by an angle γ relative to the previous adjacent control plane about the center line of the tower height, so that the tower is slightly spiral-shaped along the height direction. This biomimetic tower model is named "Beard #2". The specific parameters of the two column models are shown in Table 1.
[0037] Table 1. Dimensions of two biomimetic tower columns (unit: mm)
[0038] A B a b λ α(°) β(°) γ(°) Beard #1 13.8 5.6 11.0 6.8 21.3 15.27 17.6 0 Beard #2 101.65 51.65 88.25 61.95 265.8 57.04 61.25 2.0
[0039] The hydraulic diameter D of the seal whisker-shaped tower column was defined as the characteristic length D of the column. A traditional smooth cylindrical tower column with the same hydraulic diameter was selected for comparative research. The cylinder compared with whisker #2 was named cylinder #1.
[0040] Step 2) Using numerical simulation, the improved biomimetic tower model is compared with the cylindrical tower model. The aerodynamic characteristics of towers with different shapes are analyzed from several aspects, including surface force, velocity distribution, flow patterns, and pressure distribution. The Reynolds number R is used as the reference value. e Taking the calculation condition of 30000 and incoming wind speed U0 = 5m / s as an example, a comparative study of the surface force coefficients of the biomimetic tower column and the cylindrical tower column reveals that... Figure 3 (a) and Figure 3 As shown in (b), the lift coefficient and drag coefficient of the biomimetic tower column are both about half that of the cylindrical tower column; the statistical values of the force coefficient are shown in Table 2. The surface force fluctuation of the biomimetic tower column is significantly reduced, and it can suppress vortex-induced vibration compared with the cylindrical tower column;
[0041] Table 2. Average and root mean square of lift and drag coefficients
[0042]
[0043] Step 3) 3D print the improved biomimetic tower model and cylindrical tower model with an accuracy of 0.1mm; pressure measuring holes are evenly distributed on the nodal section and saddle point section, with 19 pressure measuring holes of 1mm in diameter in each part, so the spacing angle between the two taps is 20°;
[0044] Step 4) Set up the wind tunnel test setup and conduct the test using an electronic pressure scanning valve and a hot-wire anemometer; the specific setup can be as follows: Figure 4As shown, the device used to conduct wind tunnel comparative tests on the aerodynamic performance of the biomimetic tower column specifically includes: a biomimetic tower column model, a cylindrical tower column model, a supporting base plate, a data acquisition unit, a grid, a hot-wire anemometer, and a three-dimensional moving measurement bracket. All pressure taps on the tower model are connected to the data acquisition unit via hollow plastic hoses. A 20cm diameter disc is installed on top to eliminate interference at the ends. The data acquisition unit is the NI USB-6210 general-purpose data acquisition unit. The grid is placed at the entrance of the wind tunnel test section. The support base plate is fixed to the bottom of the wind tunnel test section, and the tower model is vertically installed on the support base plate. A 3cm opening is pre-made under the support base plate to facilitate the connection of the pressure hose to the electronic pressure acquisition system. The system has a measurement range of ±2.5kPa, a measurement accuracy of 0.05%, a sampling frequency of 333.3Hz, and a total sampling count of 10,000. The hot-wire anemometer is a DANTEC CTA / HWA constant-temperature hot-wire anemometer, placed directly behind the rotating wind turbine. The distance is adjusted in the wind tunnel using a three-dimensional moving support, which is a WNMC400.
[0045] Step 5) Use wind tunnel testing to collect data, compare the surface pressure, aerodynamic force and wake of the cylindrical tower model and the biomimetic tower model, and study the aerodynamic characteristics of towers with different shapes and structures; the surface pressure and aerodynamic force of the cylindrical tower are measured using Equation (1);
[0046]
[0047] In equation (1), p j The static pressure at test port j (j = 1~19), p0 is the incoming static pressure measured by the Pitot tube, ρ is the incoming air density, U0 is the free flow velocity, and C... p This is the pressure coefficient of the measuring hole on the cross section.
[0048] The surface pressure and aerodynamic force of the biomimetic tower column are measured using equations (2) and (3);
[0049]
[0050]
[0051] In equations (2) and (3), C l C represents the lift coefficient of each section of the biomimetic tower column. d θ is the resistance coefficient of each section of the biomimetic tower column, θ is the angle between the measuring point and the original stationary point, and its value ranges from 0° to 180°. N is the measuring point arranged within the range of 0° to 180°. In this invention, N = 10 and Δθ = θi + 1 - θi = 20° are used as an example.
[0052] The comparison results of the average pressure and standard deviation on the column surface are as follows: Figure 5 (a) and Figure 5 As shown in (b), the surface pressure distribution of the improved bionic tower is similar to that of the traditional cylindrical tower, but it has smaller surface pressure at the turning point and leeward side, and a smaller surface pressure standard deviation on the leeward side. Figure 6 The results of aerodynamic verification tests on the biomimetic tower column are presented, showing that the surface lift pulsation of the biomimetic tower column is much smaller than that of a cylindrical tower column. A comparison of the power spectral density directly behind the wake reveals that the biomimetic tower column can effectively avoid vortex-induced vibration frequencies, thus preventing vortex-induced vibration.
[0053] Step 6) Use 3D modeling software to model the wind turbine blades, hub, and nacelle, and apply the designed biomimetic tower column combined with the rotating blades to a horizontal axis wind turbine; the specific model design is as follows: Figure 7 As shown in the figure. In the specific design, the airfoil selected for the blades is DTU-LN221, the rotor radius is 0.46m, and the distance from the rotor center to the tower center is 0.15m; the tower model uses a cylindrical column and a modified biomimetic column respectively; the hub is a frustum shape, with the diameter of the upper base circle being 0.1m, the diameter of the lower base circle being 0.126m, and the generatrix being 0.072m; the nacelle is a cylinder with a diameter of 0.126m and a length of 0.35m.
[0054] Step 7) Using numerical simulation, a series of simulation results comparing the aerodynamic performance and wake field characteristics of biomimetic tower-type wind turbines and cylindrical tower-type wind turbines are studied to investigate the influence of the tower shadow effect on the load on the blades, the axial force on the rotating rotor, and the instantaneous torque for towers with different shapes; using the Reynolds number R e Taking the following calculation conditions as an example: =30000, incoming wind speed = 5m / s, rotor speed n = 300r / min, tip speed ratio λ = 3.33, the average water pressure at different heights of the cylindrical tower fluctuates around 40, which is about four times that of the biomimetic tower. The torque and thrust curves of individual blades of the biomimetic and cylindrical towers as a function of azimuth angle are shown below. Figure 8 (a) and Figure 8 As shown in (b), the thrust, torque, and power of the blades on the biomimetic tower exhibit significant periodic variations with azimuth angle. In the curves showing the variation of thrust, torque, and power with azimuth angle, the thrust, torque, and power rapidly decrease to their lowest values at azimuth angles of 60°, 180°, and 300°.
[0055] Steps 2) and 7) employ numerical simulation, using the four-equation transition model (Transition SST(4eqn)). This model is derived from γ-Re θ The transition model is coupled with the k-ωSST model, γ-Re θThe transition model combines two transport equations: transition momentum, thickness, Reynolds number, and intermittent factor. It incorporates various transition mechanisms, including natural transition, bypass transition, and separation-induced transition, offering advantages in solving complex three-dimensional fluid problems. The k-ωSST model combines the k-ε and k-ω models, accurately predicting fluid flow characteristics at the wall and adapting well to boundary conditions in the far-field region. It also includes an adverse pressure gradient correction term, making it widely applicable in adverse pressure gradient and separation flow problems such as impinging jets. This turbulence model effectively simulates the transition location and effects from laminar to turbulent flow at the model surface, improving the accuracy of calculations for complex three-dimensional fluids.
[0056] This invention demonstrates effective drag reduction and vibration suppression under various incoming flow conditions, effectively reducing aerodynamic load pulsations on the wind turbine, mitigating the tower shadow effect, suppressing vortex-induced vibration of the tower column, and improving the power output of the wind turbine. The invention also boasts low design cost, high data accuracy, and high measurement precision, significantly impacting wind energy utilization and holding significant engineering importance for wind turbine optimization research.
[0057] This invention is not limited to the above embodiments. Based on the technical solutions disclosed in this invention, those skilled in the art can make some substitutions and modifications to some of the technical features without creative effort, and all such substitutions and modifications are within the protection scope of this invention.
Claims
1. An aerodynamic design method for a biomimetic tower-type wind turbine, characterized in that, Includes the following steps: Step 1) Improve the design of the wind turbine tower using 3D modeling software. Improve the existing seal whisker model and compare it with a traditional smooth cylindrical tower with the same hydraulic diameter. The improvement design of the wind turbine tower specifically includes: changing the major and minor axis ratios and the tilt angle with the horizontal plane of the two elliptical control sections. At the same time, taking the elliptical control section at the center height of the tower as the base plane, the elliptical control sections above the base plane are rotated counterclockwise by an angle γ relative to the next adjacent control plane around the center line of the tower height. The elliptical control sections below the base plane are rotated clockwise by an angle γ relative to the previous adjacent control plane around the center line of the tower height, so that the tower is spiral in the height direction. Step 2) Using numerical simulation, the improved biomimetic tower model is compared with the cylindrical tower model. The aerodynamic characteristics of towers with different shapes and structures are studied from the aspects of surface force, velocity distribution, flow mode and pressure distribution. Step 3) 3D print the improved biomimetic tower column model and cylindrical tower column model, and evenly distribute pressure measuring holes at the nodal section and saddle point section; Step 4) Set up the wind tunnel test setup and conduct the test using an electronic pressure scanning valve and a hot-wire anemometer; Step 5) Use wind tunnel testing to collect data, compare the surface pressure, aerodynamic forces and wakes of the cylindrical tower model and the biomimetic tower model, and study the aerodynamic characteristics of towers with different shapes and structures. The wind tunnel testing method specifically includes data acquisition in the subcritical Reynolds number region where the calculated operating condition is Reynolds number R_e = 30000. The bottom of the wind tunnel testing device has a pre-set opening to facilitate the connection of a pressure hose to the electronic pressure acquisition system. A disc is installed on the top of the test column to eliminate interference at the end. Step 6) Use 3D modeling software to model the wind turbine blades, hub, and nacelle, and combine the designed biomimetic tower with the rotating blades to apply it to a horizontal axis wind turbine. Step 7) Using numerical simulation, a series of simulation results were compared between the aerodynamic performance and wake field characteristics of the biomimetic tower-type wind turbine and the cylindrical tower-type wind turbine. The study investigated the influence of the tower shadow effect on the load on the blades, the axial force on the rotating wind turbine, and the instantaneous torque for towers with different shapes and structures. The numerical simulation calculation method described in steps 2) and 7) specifically includes calculations under the Reynolds number R. e In the subcritical Reynolds number region of 30000, the applied turbulence model is based on γ-Reynolds number. θ A four-equation transition model is formed by coupling the transition model and the k-ωSST model. The simulation conditions are different wind attack angles. The biomimetic tower model and the cylindrical tower model are compared and the calculation and analysis are carried out from the aspects of surface force, velocity distribution, flow mode and pressure distribution of the tower.
2. The aerodynamic design method for a biomimetic tower-type wind turbine according to claim 1, characterized in that, The surface pressure and aerodynamic force measurement of the cylindrical tower model in step 5) is processed using equation (1); In equation (1), p j The static pressure at test port j is j = 1~19, p0 is the incoming static pressure measured by the Pitot tube, ρ is the incoming air density, U0 is the free flow velocity, and C... p This is the pressure coefficient of the measuring hole on the cross section.
3. The aerodynamic design method for a biomimetic tower-type wind turbine according to claim 1, characterized in that, The surface pressure and aerodynamic force measurement of the biomimetic tower model in step 5) are processed using equations (2) and (3); In equations (2) and (3), C l Let C be the lift coefficient of each section of the biomimetic cylinder. d θ represents the drag coefficient of each section of the biomimetic column, θ is the angle between the measuring point and the original stationary point, and its value ranges from 0° to 180°. N represents the measuring points arranged within the range of 0° to 180°.
Citation Information
Patent Citations
Wind turbine tower column imitating seal beard and design method thereof
CN109989893A
Method for optimizing aerodynamic performance of swept-curved blade wind turbine for wind tunnel experiment
CN114992044A