Ultra-high resolution natural wind field intelligent generation wind tunnel
By using an ultra-high resolution wind generation device, a high-speed turntable, and a high-frequency airfoil device, combined with deep reinforcement learning neural network technology, the problems of spatial resolution difference and turbulence characteristics in wind tunnel equipment when simulating natural wind fields have been solved, achieving efficient and intelligent wind field generation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HARBIN INSTITUTE OF TECHNOLOGY (SHENZHEN) (INSTITUTE OF SCIENCE AND TECHNOLOGY INNOVATION HARBIN INSTITUTE OF TECHNOLOGY SHENZHEN)
- Filing Date
- 2026-03-06
- Publication Date
- 2026-06-12
Smart Images
Figure CN122192686A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a wind tunnel generation method, and more particularly to an intelligent wind tunnel generation method for ultra-high resolution natural wind fields, belonging to the field of wind tunnel testing technology. Background Technology
[0002] For example, application number 201910855944.X discloses a wind tunnel simulation device and method for the sudden gust response of large port machinery and equipment. The device includes a U-shaped sudden gust simulation device formed by two guide vanes and an electric louver, a rough element scaled proportionally to the test model, a target data testing instrument, and two wind speed testing instruments respectively set at the reference position of the incoming flow and the test position. The simulation method is as follows: the sudden gust simulation device is set up in the wind tunnel and rough elements are arranged; the wind speed parameters of the simulation test are calculated; the parameters θ and ω of the electric louver are determined; the target testing equipment is connected to measure the data; and the prototype data is calculated. This wind tunnel simulation device and its testing method for the sudden gust response of large port machinery and equipment can effectively simulate the non-stationary dynamic impact effect caused by sudden gusts encountered by large port machinery and equipment, and predict the wind resistance capability of large port machinery and equipment under specific meteorological conditions in specific regions.
[0003] For example, application number 201911387663.2 discloses a connection structure for a louvered heat transfer grating, including a heat transfer body, a grating, and a connecting block. The heat transfer body is provided with a front grating surface and a rear grating surface. The grating includes a grating plate and grating strips formed by the spacing between adjacent grating plates. The heat transfer body is formed by a number of stacked upper and lower end faces and a number of front and rear attached combinations. The connecting block is horizontally shaped "H". Each heat transfer body is provided with a slot and is connected through the connecting block.
[0004] The aforementioned existing technologies still have the following problems: 1. Poor spatial resolution in wind field simulation: Existing wind tunnel equipment for simulating natural wind fields typically uses a single or dozens of large fans as wind generation units. After passing through a series of rectification and turbulence generation devices, the target wind field is created within the wind tunnel test section. The limited number of wind generation units determines the coarse spatial resolution of the generated wind field, making it impossible to simulate more refined wind fields at smaller scales. For complex terrain or urban wind fields, the spatial characteristics of natural wind can be quite complex. The poor spatial resolution of existing technologies limits their ability to accurately simulate complex natural wind fields.
[0005] 2. Accurate simulation of the turbulence characteristics of natural wind fields is difficult. Existing wind tunnel equipment for simulating natural wind fields blows air from a wind generation unit, which then passes through rectifiers such as baffles and honeycomb meshes, as well as turbulence generation devices such as rough element wedges, to generate the target wind field within the test section. The turbulence characteristics of the generated wind field depend entirely on the turbulence generation devices. Existing technical solutions are limited by the inherent characteristics of the turbulence generation devices and wind generation units, making it difficult to accurately simulate the complete turbulence characteristics of natural wind fields, especially the turbulent wind speed power spectrum and the turbulence integral scale, and also making it difficult to control the wind direction.
[0006] 3. Simulating natural wind fields involves cumbersome procedures, requiring skilled operators to spend time repeatedly debugging, and lacks automation and intelligence. Existing technical solutions require professional technicians to repeatedly debug the turbulence generation device before generating the target wind field, and the debugged device cannot be used to simulate other target wind fields. This results in time-consuming and labor-intensive wind field debugging and a serious lack of intelligence.
[0007] Based on the above shortcomings, an ultra-high resolution natural wind field intelligent generation wind tunnel is needed to optimize these shortcomings. Summary of the Invention
[0008] The main objective of this invention is to overcome: 1. Poor spatial resolution in wind field simulation: Existing wind tunnel equipment for simulating natural wind fields typically uses a single or dozens of large fans as wind generation units. After passing through a series of rectification and turbulence generation devices, the target wind field is created within the wind tunnel test section. The limited number of wind generation units determines the coarse spatial resolution of the generated wind field, making it impossible to simulate more refined wind fields at smaller scales. For complex terrain or urban wind fields, the spatial characteristics of natural wind can be quite complex. The poor spatial resolution of existing technologies limits their ability to accurately simulate complex natural wind fields.
[0009] 2. Accurate simulation of the turbulence characteristics of natural wind fields is difficult. Existing wind tunnel equipment for simulating natural wind fields blows air from a wind generation unit, which then passes through rectifiers such as baffles and honeycomb meshes, as well as turbulence generation devices such as rough element wedges, to generate the target wind field within the test section. The turbulence characteristics of the generated wind field depend entirely on the turbulence generation devices. Existing technical solutions are limited by the inherent characteristics of the turbulence generation devices and wind generation units, making it difficult to accurately simulate the complete turbulence characteristics of natural wind fields, especially the turbulent wind speed power spectrum and the turbulence integral scale, and also making it difficult to control the wind direction.
[0010] 3- Simulating natural wind fields involves cumbersome procedures, requiring skilled operators to spend time repeatedly debugging, lacking automation and intelligence. Existing technical solutions require professional technicians to repeatedly debug the turbulence generation device before generating the target wind field, and the debugged device cannot be used to simulate other target wind fields. This results in time-consuming and labor-intensive wind field debugging and a serious lack of intelligence, necessitating the provision of an ultra-high resolution intelligent wind tunnel for generating natural wind fields.
[0011] The objective of this invention can be achieved by adopting the following technical solution: An ultra-high resolution natural wind field intelligent generation wind tunnel, including a natural wind generation field area for simulation; An ultra-high resolution wind production device is installed at one end of the natural wind generation area, a high-frequency wing plate device is installed on one side of the ultra-high resolution wind production device, and high-speed turntables are symmetrically installed at the bottom of the natural wind generation area. A sensor array is installed on one side of the natural wind generation area.
[0012] Preferably, the height of the ultra-high resolution wind production device is 0.7m-2m and the width is 1.4m-2.6m.
[0013] Preferably, the ultra-high resolution wind production device has 16-24 rows horizontally and 24-36 columns vertically.
[0014] Preferably, a fixed outer frame is installed on the outside of the ultra-high resolution wind production device, and high-frequency wing plate devices are distributed on the inside of the ultra-high resolution wind production device.
[0015] Preferably, the high-frequency wingplate device has wingplates distributed at an angle.
[0016] Preferably, a reduction gear set is installed at the bottom of the high-speed turntable, and the reduction gear set and the high-speed turntable are driven by a high-speed drive motor.
[0017] Preferably, the natural wind generation area is provided with two sets of high-speed turntables, high-speed gear sets and high-speed drive motors.
[0018] Preferably, the upper wing plate of the high-frequency wing plate device is a movable structure.
[0019] Beneficial technical effects of the present invention: The present invention provides an ultra-high resolution intelligent wind tunnel for generating natural wind fields. A rectangular space is left on the front of the fixed outer frame for the fan array to be embedded within. The rectangular frame at the front of the ultra-high resolution fan array serves as the test section of the device. The cross-sectional dimensions of the test section are consistent with those of the ultra-high resolution fan array. The fan array as a whole is supported and fixed by the fixed outer frame 101. The outer frame is directly erected and anchored to the ground. The back side of the outer frame is reinforced with counterweights and diagonal braces to enhance the gravity balance and stability of the wind wall device. The fan array is matrix-shaped and has variable dimensions and fan specifications. Dimensions include 2 meters high and 2.6 meters wide, 0.7 meters high and 1.4 meters wide, etc. Specifications include 24 rows and 32 columns of ultra-high wind rate fan arrays, 16 rows and 24 columns of fan arrays, etc. The specifications and dimensions of the ultra-high wind rate fan array can be customized. The size and other technical parameters of the ultra-high wind rate fan array can also be customized. Therefore, this ultra-high resolution wind generation device is a unique technical solution rather than a specific implementation method. Compared with traditional multi-fan wind generation devices, this ultra-high resolution wind generation device has a much denser ultra-high wind rate fan array per unit area, thus generating a more refined unit wind field with ultra-high wind field resolution. Each ultra-high wind rate fan array is essentially a minimum wind generation unit. With the power supply system and the control system described below, arbitrary unidirectional wind time history can be customized to generate. In terms of mechanical principle, the rear of the high-efficiency fan array absorbs air and creates an air pressure difference by rotating the blades through an internal motor, thereby generating airflow in front of the high-efficiency fan array. Inside the fan array, the individual high-efficiency fan arrays are further fixed and integrated by fasteners. Each high-efficiency fan array can be disassembled or connected individually. A total of 16 high-efficiency fan arrays in 4 rows and 4 columns are connected to the same PCB unit. This PCB unit is responsible for the power supply and control signal transmission of the 16 high-efficiency fan arrays below it. The high-frequency wing plate device is set between the ultra-high resolution wind wall device 1 and the test section, close to the outer frame of the wind wall array. Here, the high-frequency wing plate device is equipped with 16 independent rotating wing plates, each driven by a drive motor. The rotation of the wing plates adds disturbance to the airflow blown out by the fan array. In the middle and later sections of the test section, a simulated natural wind field is formed. The high-speed drive motor is the power source of the high-speed turntable device. The high-speed rotation of the motor can be precisely controlled by electrical signals. The reduction gear set serves as a transmission device to transmit the power output from the motor shaft to the high-speed turntable fixing frame. The fixing frame connects the upper high-speed turntable with the lower acceleration gear set, transmitting the power of the gear set to the high-speed turntable device to drive its rotation. A sensor array is set up at the target wind field location to collect multi-dimensional spatiotemporal information such as wind speed, wind pressure, and temperature, which serves as the input end of the automatic wind field generation algorithm. Once the algorithm is trained, it can realize the automatic feedback adjustment and generation of the target wind field without manual debugging. The ultra-high resolution wind production device includes a fan array composed of a large number of regularly arranged ultra-high wind rate fan arrays and corresponding power supply and control systems. In terms of control principle, the ultra-high wind rate fan array receives pulse width modulation (PWM) signals. The PWM signals determine the equivalent operating voltage of the motor, thereby controlling the rotation speed of the blades. Thus, the wind speed of the airflow generated in front of the ultra-high wind rate fan array is controlled. By controlling the real-time changing PWM signals given by the control system, the airflow in front of each ultra-high wind rate fan array can be generated in real time, which can customize the generated wind time history. This ultra-high resolution wind generation device is essentially a collection of all ultra-high wind rate fan arrays in the fan array. Since each ultra-high wind rate fan array in this array can be independently controlled to generate a specific wind time history, a more complex and refined spatiotemporal wind field can be generated within the coverage area of the fan array compared to traditional technical solutions. The high-frequency airfoil device is one or more high-frequency airfoil devices stacked in different directions. The device includes a row of parallel rotating blades with the same plane of rotation axis, an outer frame for fixing these blades, and a drive motor system for controlling the rotation of the blades. Since the ultra-high resolution wind generating device shown in the ultra-high resolution wind production device can only generate unidirectional wind along the fan rotation axis and lacks the ability to generate multidirectional wind, the high-frequency airfoil device shown in the high-frequency airfoil device actually functions to disrupt the unidirectional wind field generated by the ultra-high resolution wind generating device and create wind speed and air volume in other directions in the wind field, thereby providing more dimensions of turbulence characteristic adjustment capability. In terms of mechanical principle, the rotating blade has a streamlined or other cross-sectional shape, and its angle around the axis can be adjusted in real time as needed, or it can be in continuous reciprocating rotation around the axis. In terms of control principle, each rotating blade is driven by an independent drive motor, which can arbitrarily control the blade to maintain a specific angle around the axis, or control its rotational movement in real time, and adjust the rotation angle range and speed. The generated unidirectional wind produces complex airflow disturbances when it flows through the blades which are in a complex motion state. After flowing through the blades, it no longer maintains unidirectional wind, but has wind speed components in multiple directions. Furthermore, since the unidirectional wind generated by the ultra-high resolution wind production device has complex spatiotemporal characteristics, the wind field further exhibits complex turbulent characteristics after passing through the high-frequency wing plate device. The significance of the high-frequency airfoil device lies in its ability to greatly expand the turbulence characteristics that the wind field generated by ultra-high resolution wind production devices can possess. The intelligent controller based on deep reinforcement learning neural network technology takes the complex natural wind field to be simulated as the control target. By collecting the turbulence characteristics of different spatial locations of the wind field in the test section, inputting them into the deep neural network, and outputting control commands to various hardware components, after an automated training process, it can realize the intelligent simulation of complex natural wind fields without the need for manual adjustment of physical parameters such as the adjustable turbulence intensity range and turbulence integral scale range in high wind fields.
[0020] The high-speed turntable consists of one or more high-speed turntable devices. The ultra-high resolution wind production device and the high-frequency wing plate device jointly generate an ultra-high resolution simulated natural wind field to provide an experimental environment for wind tunnel experiments. The test object, such as a building model, needs to be fixedly placed on the high-speed turntable. The upper part of the high-speed turntable is flush with the bottom surface of the wind tunnel test section. The model or equipment to be tested can be fixed on the high-speed turntable. The wind direction angle of the model is changed by rotating the high-speed turntable. The high-speed turntable provides power and control through the drive motor and reduction gear system below it. It has a finely controllable speed and direction, so it can provide a real-time changing wind direction angle for the model fixed on the high-speed turntable. The intelligent controller based on deep reinforcement learning neural network technology takes the complex natural wind field to be simulated as the control target. By collecting turbulence characteristics at different spatial locations of the wind field in the test section, inputting them into the deep neural network, and outputting control commands to each hardware component, it can realize intelligent simulation of complex natural wind fields after an automated training process without manual debugging. This intelligent control algorithm is mainly based on deep reinforcement learning neural network technology. The characteristic of deep reinforcement learning technology is that it does not require algorithm settings and adjustments for complex and specific control parameters and interactive environments. Instead, it uses interactive behavior control methods such as designing network structure, network hyperparameters, and reward functions to enable the neural network agent to work autonomously in the interactive space composed of complex control variables, to evaluate its control effect and optimize itself, and to continuously iterate the training process. Ultimately, it can quickly find the optimal parameters in the control space without human intervention. The advantage of this intelligent control algorithm is that it does not require prior knowledge of how the control parameters will ultimately affect the control effect. This is because how the ultra-high resolution wind production device, high-frequency wing plate device, and high-speed turntable hardware jointly affect the generation of the final ultra-high resolution simulated natural wind field is a complex fluid dynamic process, and such prior knowledge is relatively lacking. The intelligent control algorithm takes discrete wind field data collected by multiple sensor arrays set up in the wind field as its input, including data such as wind speed, turbulence intensity, wind pressure, and temperature in multiple directions at each measuring point. The output of the algorithm is the control parameters of the ultra-high resolution wind production device, the high-frequency airfoil device, and the high-speed turntable device, including the pulse width modulation signals of different fans in the ultra-high resolution wind production device, the rotation angle and rotation mode of different blades in the high-frequency airfoil device, and the rotation angle and direction of the high-speed turntable in the high-speed turntable. The training process uses a large number of given complex target wind field conditions as reference values. The training process is as follows: After the intelligent control algorithm outputs control parameters, multiple sensors installed in the wind field collect various parameters of the wind field generated under the current control strategy. The built-in reward function and other mechanisms will automatically evaluate the difference between the generated wind field and the target wind field and adjust the control strategy accordingly, thereby updating the network parameters and outputting a new round of control parameters. The network parameter updates and iterations converge during the extensive training process, and the difference between the wind field generated by the control algorithm and the target wind field is reduced to within the allowable error range for engineering. After training is completed, the intelligent control algorithm will automatically simulate other complex natural wind fields as required. Attached Figure Description
[0021] Figure 1 This is a schematic diagram of the hardware composition of a preferred embodiment of the ultra-high resolution natural wind field intelligent generation wind tunnel according to the present invention; Figure 2 A diagram of an ultra-high resolution wind generation device according to a preferred embodiment of an ultra-high resolution natural wind field intelligent generation wind tunnel of the present invention; Figure 3 A high-frequency airfoil device diagram of a preferred embodiment of the intelligent generation wind tunnel for ultra-high resolution natural wind fields according to the present invention; Figure 4 A diagram of a high-speed turntable 3 device according to a preferred embodiment of an ultra-high resolution natural wind field intelligent generation wind tunnel of the present invention; Figure 5 This is a schematic diagram of the sensor setup for the automatic wind field generation algorithm according to a preferred embodiment of the intelligent wind tunnel for generating ultra-high resolution natural wind fields of the present invention. Figure 6 This is a schematic diagram illustrating the implementation principle of the intelligent control software algorithm in a preferred embodiment of the intelligent generation wind tunnel for ultra-high resolution natural wind fields according to the present invention. Figure 7 This is a schematic diagram of a preferred embodiment of an ultra-high resolution wind generation device according to the present invention, which is a smart wind tunnel for generating ultra-high resolution natural wind fields.
[0022] In the figure: 1. Ultra-high resolution wind generator; 101. Fixed outer frame; 102. High wind rate fan array; 2. High-frequency wingplate device; 201. Wingplate; 3. High-speed turntable 3; 301. Reduction gear set; 302. High-speed drive motor; 4. Natural wind generation area; 401. Sensor array. Detailed Implementation
[0023] To enable those skilled in the art to understand the technical solution of the present invention more clearly, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.
[0024] Example 1: As Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 As shown, this embodiment proposes an ultra-high resolution natural wind field intelligent generation wind tunnel, including a natural wind generation field region 4 for simulation. An ultra-high resolution wind production device 1 is installed at one end of the natural wind generation field area 4, a high frequency wing plate device 2 is installed on one side of the ultra-high resolution wind production device 1, and a high speed turntable 3 is symmetrically installed at the bottom of the natural wind generation field area 4. A sensor array 401 is installed on one side of the natural wind generation field area 4.
[0025] The ultra-high resolution wind production device 1 has a height of 0.7m-2m and a width of 1.4m-2.6m.
[0026] The ultra-high resolution wind production device 1 has 16-24 rows horizontally and 24-36 columns vertically.
[0027] The ultra-high resolution wind production device 1 is equipped with a fixed outer frame 101 on its outer side, and high-frequency wing plate devices 2 are distributed on the inner side of the ultra-high resolution wind production device 1.
[0028] The high-frequency wing plate device 2 has wing plates 201 distributed obliquely on it; like Figure 1 and Figure 3 As shown, a rectangular space is left on the front of the fixed outer frame 101 for the fan array to be embedded in. The rectangular frame set at the front end of the ultra-high resolution fan array is the test section of the device. The cross-sectional dimensions of the test section are consistent with those of the ultra-high resolution fan array. The middle and rear section of the test section is the natural wind field simulation area 4 shown in Figure 4. The fan array as a whole is supported and fixed by the fixed outer frame 101. The outer frame is directly erected and anchored to the ground. The gravity balance and stability of the wind wall device are enhanced by counterweights and diagonal braces on the back side of the outer frame. The fan array is matrix-shaped and has variable dimensions and fan specifications, such as 2 meters high and 2.6 meters wide, 0.7 meters high and 1.4 meters wide, etc., and specifications such as 24 rows and 32 columns. The specifications and dimensions of the ultra-high wind rate fan array, such as the 16-row 24-column ultra-high wind rate fan array, can be customized. The size and other technical parameters of the ultra-high wind rate fan array can also be customized. Therefore, this ultra-high resolution wind generation device 1 is a unique technical solution rather than a specific implementation method. Compared with traditional multi-fan wind generation devices, this ultra-high resolution wind generation device 1 has a much denser ultra-high wind rate fan array per unit area, thus generating a more refined unit wind field with ultra-high wind field resolution. Each ultra-high wind rate fan array is essentially a minimum wind generation unit. With the power supply system and the control system described below, arbitrary unidirectional wind time history can be customized to generate. In terms of mechanical principle, the rear of the high-efficiency fan array absorbs air and creates an air pressure difference by rotating the blades through an internal motor, thereby generating airflow in front of the high-efficiency fan array. Inside the fan array, the individual high-efficiency fan arrays are further fixed and integrated by fasteners. Each high-efficiency fan array can be disassembled or connected individually. A total of 16 high-efficiency fan arrays in 4 rows and 4 columns are connected to the same PCB unit. This PCB unit is responsible for the power supply and control signal transmission of the 16 high-efficiency fan arrays below it. The high-frequency wing plate device 2 is set between the ultra-high resolution wind wall device 1 and the test section, close to the outer frame of the wind wall array. Here, the high-frequency wing plate device 2 is equipped with 16 independent rotating wing plates, which are driven by drive motors 302 respectively. The rotation of the wing plates 201 adds disturbance to the airflow blown out by the fan array. In the middle and later sections of the test section, a simulated natural wind field is formed. The high-speed drive motor 302 is the power source of the high-speed turntable 3. The high-speed rotation of the motor can be precisely controlled by electrical signals. The reduction gear set 1 serves as a transmission device to transmit the power output from the motor shaft to the fixed frame of the high-speed turntable 3. The fixed frame connects the upper high-speed turntable 3 with the lower acceleration gear set, transmitting the power of the gear set to the high-speed turntable 3 to drive its rotation. A sensor array 401 is set up at the target wind field location to collect multi-dimensional spatiotemporal information such as wind speed, wind pressure, and temperature, which serves as the input end of the wind field automatic generation algorithm. Once the algorithm is trained, it can realize the automatic feedback adjustment and generation of the target wind field without manual debugging. The ultra-high resolution wind production device 1 includes a fan array composed of a large number of regularly arranged ultra-high wind rate fan arrays 102 and a corresponding power supply and control system. In terms of control principle, the ultra-high wind rate fan array 102 receives a pulse width modulation (PWM) signal. The PWM signal determines the equivalent operating voltage of the motor, thereby controlling the rotation speed of the blades. Thus, it controls the wind speed of the airflow generated in front of the ultra-high wind rate fan array 102. By controlling the real-time changing PWM signal given by the control system, the airflow in front of each ultra-high wind rate fan array can be generated in real time, and the wind time course can be customized. The ultra-high resolution wind generation device 1 is essentially a collection of all ultra-high wind rate fan arrays 102 in the fan array. Since each ultra-high wind rate fan array 102 in the array can be independently controlled to generate a specific wind time course, a more complex and refined spatiotemporal wind field can be generated within the coverage area of the fan array compared with traditional technical solutions. The high-frequency airfoil device 2 is one or more high-frequency airfoil devices 2 stacked in different directions. The device includes a row of parallel rotating blades with the same plane of rotation axis, an outer frame for fixing these blades, and a drive motor 302 system for controlling the rotation of the blades, etc. Since the ultra-high resolution wind generating device 1 shown in the ultra-high resolution wind generating device 1 can only generate unidirectional wind along the fan rotation axis and lacks the ability to generate multidirectional wind, the high-frequency air vane device 2 shown in the high-frequency air vane device 2 actually functions to disrupt the unidirectional wind field generated by the ultra-high resolution wind generating device 1 and generate wind speed and air volume in other directions in the wind field, thereby providing more dimensions of turbulence characteristic adjustment capability. In terms of mechanical principle, the rotating blade has a streamlined or other cross-sectional shape, and its angle around the axis can be adjusted in real time as needed, or it can be in continuous reciprocating rotation around the axis. In terms of control principle, each rotating blade is driven by an independent drive motor 302, which can arbitrarily control the blade to maintain a specific angle around the axis, or control its rotational movement in real time, and adjust the rotation angle range and speed. When the generated unidirectional wind flows through the blades in a complex motion state, it generates complex airflow disturbances. After flowing through the blades, it no longer maintains unidirectional wind, but has wind speed components in multiple directions. Furthermore, since the unidirectional wind generated by the ultra-high resolution wind production device 1 has complex spatiotemporal characteristics, after passing through the high-frequency wing plate device 2, the wind field further exhibits complex turbulent characteristics. The significance of the high-frequency airfoil device 2 lies in its ability to greatly expand the turbulence characteristics of the wind field generated by the ultra-high resolution wind production device 1. The intelligent controller based on deep reinforcement learning neural network technology takes the complex natural wind field to be simulated as the control target. By collecting the turbulence characteristics of different spatial locations of the wind field in the test section, inputting them into the deep neural network, and outputting control commands to each hardware component, the intelligent controller can realize the intelligent simulation of complex natural wind fields after an automated training process. There is no need to manually adjust the adjustable turbulence intensity range, turbulence integral scale range, and other physical parameters in high wind fields.
[0029] Example 2: The solution in Example 1 will be further described below with reference to its specific working method. See the description below for details: like Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 and Figure 7 As shown, in a preferred embodiment, based on the above method, a reduction gear set 301 is further installed at the bottom of the high-speed turntable 3, and the reduction gear set 301 and the high-speed turntable 3 are driven by a high-speed drive motor 302.
[0030] The natural wind generation area 4 contains two sets of high-speed turntables 3, a high-speed gear set 301, and a high-speed drive motor 302.
[0031] The upper wing plate 201 of the high-frequency wing plate device 2 is a movable structure.
[0032] like Figure 1 , Figure 2 , Figure 3 and Figure 4 As shown, the high-speed turntable 3 is one or more high-speed turntable 3 devices. The ultra-high resolution wind production device 1 and the high-frequency wing plate device 2 jointly generate an ultra-high resolution simulated natural wind field to provide an experimental environment for wind tunnel experiments. The experimental objects, such as architectural models, need to be fixedly placed on the high-speed turntable 3. The high-speed rotation of the motor 302 can be precisely driven by electrical signals. The reduction gear set 301 serves as a transmission device to transmit the power output from the motor shaft to the turntable fixing frame. The fixing frame connects the upper turntable and the lower acceleration gear set, transmitting the power of the gear set to the high-speed turntable 3 to drive its rotation. The upper part of the high-speed turntable 3 is flush with the bottom surface of the wind tunnel test section. The model or equipment to be tested can be fixed on the high-speed turntable 3. The wind direction angle of the model can be changed by rotating the high-speed turntable 3. The high-speed turntable 3 provides power and control through the drive motor 302 and reduction gear set 301 system below it. It has a finely controllable speed and direction, so it can provide a real-time changing wind direction angle for the model fixed on the high-speed turntable 3. The intelligent controller based on deep reinforcement learning neural network technology takes the complex natural wind field to be simulated as the control target. By collecting turbulence characteristics at different spatial locations of the wind field in the test section, inputting them into the deep neural network, and outputting control commands to each hardware component, it can realize intelligent simulation of complex natural wind fields after an automated training process without manual debugging. This intelligent control algorithm is mainly based on deep reinforcement learning neural network technology. The characteristic of deep reinforcement learning technology is that it does not require algorithm settings and adjustments for complex and specific control parameters and interactive environments. Instead, it enables the neural network agent to work in the interactive space composed of complex control variables by designing network structures, network hyperparameters and reward functions, evaluate its control effect and optimize itself, and continuously iterate the training process. Finally, it can quickly find the optimal parameters in the control space without human intervention.
[0033] The advantage of this intelligent control algorithm is that it does not require prior knowledge of how the control parameters will ultimately affect the control effect. This is because the hardware devices of the ultra-high resolution wind production device 1, the high-frequency wing plate device 2, and the high-speed turntable 3 in this scheme jointly affect the generation of the final ultra-high resolution simulated natural wind field, which is a complex fluid dynamic process, and such prior knowledge is relatively lacking. The intelligent control algorithm uses discrete wind field data collected by multiple sensor arrays 401 installed in the wind field as input, including data such as wind speed, turbulence intensity, wind pressure, and temperature in multiple directions at each measuring point. The output of the algorithm is the control parameters of the ultra-high resolution wind production device 1, the high-frequency blade device 2, and the high-speed turntable 3, including the pulse width modulation signals of different fans in the ultra-high resolution wind production device 1, the rotation angle and rotation shape of different blades in the high-frequency blade device 2, and the rotation angle and direction of the high-speed turntable 3. The training process uses a large number of given complex target wind field conditions as reference values. The training process is as follows: After the intelligent control algorithm outputs control parameters, multiple sensors installed in the wind field collect various parameters of the wind field generated under the current control strategy. The built-in reward function and other mechanisms will automatically evaluate the difference between the generated wind field and the target wind field and adjust the control strategy accordingly, thereby updating the network parameters and outputting a new round of control parameters. The network parameter updates and iterations converge during the extensive training process, and the difference between the wind field generated by the control algorithm and the target wind field is reduced to within the allowable error range for engineering. After training is completed, the intelligent control algorithm will automatically simulate other complex natural wind fields as required.
[0034] The above description is merely a further embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope disclosed in the present invention, based on the technical solution and concept of the present invention, shall fall within the scope of protection of the present invention.
Claims
1. A high-resolution natural wind field intelligent generation wind tunnel, including a natural wind generation field area for simulation (4). Its features are: An ultra-high resolution wind production device (1) is installed at one end of the natural wind generation field area (4), a high frequency wing plate device (2) is installed on one side of the ultra-high resolution wind production device (1), and a high speed turntable (3) is symmetrically installed at the bottom of the natural wind generation field area (4). A sensor array (401) is installed on one side of the natural wind generation field area (4).
2. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 1, characterized in that: The ultra-high resolution wind production device (1) has a height of 0.7m-2m and a width of 1.4m-2.6m.
3. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 2, characterized in that: The ultra-high resolution wind production device (1) has 16-24 rows horizontally and 24-36 columns vertically.
4. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 3, characterized in that: The ultra-high resolution wind production device (1) is equipped with a fixed outer frame (101) on the outside and a high-frequency wing plate device (2) is distributed on the inside of the ultra-high resolution wind production device (1).
5. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 4, characterized in that: The high-frequency wing plate device (2) has wing plates (201) distributed obliquely on it.
6. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 1, characterized in that: The high-speed turntable (3) is equipped with a reduction gear set (301) at its bottom, and the reduction gear set (301) and the high-speed turntable (3) are driven by a high-speed drive motor (302).
7. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 6, characterized in that: The natural wind generation field area (4) is equipped with two sets of high-speed turntables (3), a speed gear set (301), and a high-speed drive motor (302).
8. The ultra-high resolution natural wind field intelligent generation wind tunnel according to claim 5, characterized in that: The upper wing plate (201) of the high-frequency wing plate device (2) is a movable structure.