Multi-fan active control direct current blowing type micro wind tunnel and wind speed profile debugging method thereof

Abstract

The invention discloses a multi-fan active control direct current blowing type micro wind tunnel and a wind speed profile debugging method thereof, the multi-fan active control direct current blowing type micro wind tunnel comprises a micro wind tunnel, a multi-fan array, a fan control system and a wind speed measuring system, the multi-fan array is arranged in the micro wind tunnel, and the multi-fan array comprises a plurality of fans; the fan control system is electrically connected to the multi-fan array and used for controlling the rotating speed of the multiple fans. The wind speed measuring system is used for obtaining the wind speed in the micro wind tunnel. According to the multi-fan active control direct current blowing type micro wind tunnel, the rotating speed of each fan is controlled through the fan control system, so that multi-dimensional wind speed profile surface distribution is achieved, and the flexibility and precision of wind speed control are greatly improved.

Description

Technical Field

[0001] The invention relates to the field of fluid mechanics and aerodynamic experimental technology, and in particular to a multi-fan active-control direct-current blowing micro wind tunnel and a wind speed profile debugging method thereof. Background Art

[0002] Among the related technologies, wind tunnel tests are becoming increasingly important in aerodynamics, fluid mechanics and engineering testing, especially in areas such as aircraft, buildings and wind turbines that have strict requirements on airflow behavior. Accurate simulation of the fluid environment is crucial. Traditional wind tunnels mostly use a single fan or a small number of fans to drive the fluid. This method has limitations in flow rate control, especially it is difficult to achieve complex wind speed profile distribution and subtle flow rate adjustment. Accurate control of the wind speed profile in the wind tunnel helps to simulate complex airflow conditions and provide a reliable experimental basis for scientific research and engineering design.

[0003] The need to control the wind speed profile is particularly important in micro wind tunnels. Micro wind tunnels are used for small-scale experiments, such as micro UAVs, design and testing of aerodynamic components. Due to the small experimental section of the micro wind tunnel, the traditional single-fan DC blowing structure is difficult to obtain an ideal uniform wind speed field, and the adjustment space between each wind speed point is also limited. In recent years, the scheme of achieving flow field control by adding a multi-fan array structure has received much attention. Multi-fan arrays can refine the flow field control, but the problem of controlling the wind speed profile is still relatively complex, especially under dynamic changes or multi-objective requirements. It is still difficult to independently control each fan. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, one purpose of the present invention is to propose a multi-fan active control DC blowing micro wind tunnel, which controls the speed of each fan through a fan control system to achieve multi-dimensional wind speed profile distribution, greatly improving the flexibility and accuracy of wind speed control.

[0005] The present invention also provides a wind speed profile debugging method for the above-mentioned multi-fan active control DC blowing micro wind tunnel.

[0006] According to the first aspect of the present invention, a multi-fan actively controlled DC blowing micro wind tunnel includes: a micro wind tunnel; a multi-fan array, wherein the multi-fan array is arranged in the micro wind tunnel, and the multi-fan array includes a plurality of fans; a fan control system, wherein the fan control system is electrically connected to the multi-fan array and is used to control the rotation speed of the fans; and a wind speed measurement system, wherein the wind speed measurement system is used to obtain the wind speed in the micro wind tunnel.

[0007] According to the multi-fan active DC blowing micro wind tunnel of the embodiment of the present invention, the rotation speed of each fan is controlled by a fan control system to achieve multi-dimensional wind speed profile distribution, thereby greatly improving the flexibility and accuracy of wind speed control.

[0008] According to some embodiments of the present invention, in the direction from the air flow inlet of the micro wind tunnel to the air flow outlet of the micro wind tunnel, the cross-section of the wind tunnel is square, the micro wind tunnel includes a power section, a contraction section, an experimental section and a diffusion section connected in sequence, the multi-fan array is arranged in the power section, and the experimental section is a transparent part that can transmit light.

[0009] According to some embodiments of the present invention, the multi-fan array is located in the power section, the number of the fans is an even square number, and the multiple fans are arranged in a matrix and abut against each other in sequence.

[0010] According to some embodiments of the present invention, the contraction section is provided with a plurality of internal channels, the number of the internal channels is equal to and corresponds one to one with the fans, the boundary curve of the internal channel is designed according to the Vidosinski curve, and a honeycomb is provided at the end of the contraction section, and the honeycomb is detachably connected to the micro wind tunnel.

[0011] According to some embodiments of the present invention, a multi-fan actively controlled DC blowing micro wind tunnel further includes: a computer system, the fan control system includes: a pulse width modulation drive control board and a single chip microcomputer, the system control board is electrically connected to the single chip microcomputer, the fan is connected to the single chip microcomputer, and the computer system controls the PWM duty cycle of each of the fans through the pulse width modulation drive control board and the single chip microcomputer to adjust the rotation speed of each of the fans.

[0012] According to some embodiments of the present invention, the multi-fan actively controlled DC blowing micro wind tunnel also includes: a computer system, the fan control system is electrically connected to the computer system, the wind speed measurement system includes: a three-axis movable measuring frame and a hot-wire anemometer, the hot-wire anemometer includes a sensor probe, the three-axis movable measuring frame is installed in the experimental section, the hot-wire anemometer is located outside the wind tunnel, the sensor probe extends into the experimental section and is installed on the three-axis movable measuring frame, and the computer system is used to adjust the position of the three-axis movable measuring frame.

[0013] According to the wind speed profile debugging method of the second embodiment of the present invention, it is applied to the multi-fan active control DC blowing micro wind tunnel of the first embodiment of the present invention. The wind speed profile debugging method includes a convolutional neural network and a particle swarm optimization algorithm. The PWM duty cycle of each fan is adjusted based on the convolutional neural network and the particle swarm optimization algorithm to achieve the target wind speed profile.

[0014] According to the wind speed profile debugging method of an embodiment of the present invention, a convolutional neural network and a particle swarm optimization algorithm are used to accurately control the PWM duty cycle of a multi-fan array to achieve a target wind speed profile.

[0015] According to some embodiments of the present invention, in the direction from the air flow inlet of the micro wind tunnel to the air flow outlet of the micro wind tunnel, the micro wind tunnel includes a power section, a contraction section, an experimental section and a diffusion section connected in sequence, the convolutional neural network is used to predict the wind speed profile of the multi-fan array in the experimental section at different duty cycles, and the duty cycle scheme of the multi-fan array that best matches the target wind speed profile is determined according to the particle swarm optimization algorithm.

[0016] According to some embodiments of the present invention, the convolutional neural network includes an encoder and a decoder, the encoder is used to receive a two-dimensional array of the multi-fan array at different duty cycles, and the decoder generates a normalized two-dimensional array of downstream average wind speed on a cross-sectional point array within the experimental section.

[0017] 10. According to some embodiments of the present invention, the convolutional neural network is trained using training data, wherein the training data is composed of measurement results of the average downstream wind speed on the cross-sectional lattice of the experimental section under different speed schemes of the multi-fan array, and the measurement results are normalized based on the maximum average downstream wind speed on the cross-sectional lattice of the experimental section.

[0018] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0020] Figure 1 is a schematic diagram of a multi-fan active control direct current blowing micro wind tunnel according to some embodiments of the present invention;

[0021] Figure 2 Schematic diagram of the structure of a convolutional neural network according to an embodiment of the present invention.

[0022] Reference numerals:

[0023] 10. Multiple fans actively control the DC blowing breeze tunnel;

[0024] 1. Micro wind tunnel; 11. Power section; 12. Contraction section; 121. Honeycomb; 13. Experimental section; 14. Diffusion section;

[0025] 2. Multi-fan array;

[0026] 3. Fan control system; 31. Fan power supply; 32. Pulse width modulation drive control board; 33. Single chip microcomputer;

[0027] 4. Wind speed measurement system; 41. Three-axis mobile measuring frame; 42. Hot wire anemometer; 43. Sensor probe;

[0028] 5. Computer system. DETAILED DESCRIPTION

[0029] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0030] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0031] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0032] Reference below Figure 1-Figure 2 A multi-fan active control direct current blowing micro wind tunnel 10 according to an embodiment of the present invention is described.

[0033] According to the multi-fan actively controlled DC blowing micro wind tunnel 10 of the first aspect of the present invention, the multi-fan actively controlled DC blowing micro wind tunnel 10 includes a micro wind tunnel 1, a multi-fan array 2, a fan control system 3 and a wind speed measurement system 4, the multi-fan array 2 is arranged in the micro wind tunnel 1; the fan control system 3 is electrically connected to the multi-fan array 2, the multi-fan array 2 includes a plurality of fans, and is used to control the rotation speed of the plurality of fans; the wind speed measurement system 4 is used to obtain the wind speed in the micro wind tunnel 1.

[0034] The fan control system 3 can control the rotation speed of each fan and obtain the wind speed of multiple fans at different rotation speeds in the micro wind tunnel 1 through the wind speed measurement system 4 to achieve multi-dimensional wind speed profile distribution, greatly improving the flexibility and accuracy of wind speed control.

[0035] According to the multi-fan active DC blowing micro wind tunnel 10 of the embodiment of the present invention, the rotation speed of each fan is controlled by the fan control system 3 to achieve multi-dimensional wind speed profile distribution, thereby greatly improving the flexibility and accuracy of wind speed control.

[0036] According to some embodiments of the present invention, referring to Figure 1 In the direction from the airflow inlet of the breeze tunnel 1 to the airflow outlet of the breeze tunnel 1, the breeze tunnel 1 includes a power section 11, a contraction section 12, an experimental section 13 and a diffusion section 14 connected in sequence. In the direction from the airflow inlet of the breeze tunnel 1 to the airflow outlet of the breeze tunnel 1, the cross-sectional area of ​​the contraction section 12 gradually decreases, the cross-sectional area of ​​the experimental section 13 gradually remains unchanged, and the cross-sectional area of ​​the diffusion section 14 gradually increases. The power section 11 is provided with a multi-fan array 2 as a wind source to generate a stable airflow; the contraction section 12 is used to stabilize and accelerate the airflow, the experimental section 13 is an area for observing the airflow characteristics, and the diffusion section 14 is used to discharge the airflow smoothly.

[0037] The experimental section 13 is a transparent part that can transmit light. The experimental section 13 is made of transparent organic glass, which is conducive to directly observing the experimental phenomenon and facilitating the use of external optical equipment. In the direction from the airflow inlet of the micro-wind tunnel 1 to the airflow outlet of the micro-wind tunnel 1, the cross-section of the micro-wind tunnel 1 is square, and each section body adopts a square cross-section design to simplify the fluid flow path and minimize interference.

[0038] For example, in the direction from the airflow inlet of the breeze tunnel 1 to the airflow outlet of the breeze tunnel 1 , the length and width of the cross section of the experimental section 13 are both 0.72 m.

[0039] According to some embodiments of the present invention, referring to Figure 1 The number of fans is an even square number, and multiple fans are arranged in a square matrix and are set in sequence, which can effectively avoid flow interference in the gaps between fans, reduce the airflow boundary layer effect, and ensure the uniformity of the airflow.

[0040] Specifically, N fans are set in each row and N fans are set in each column, where N is an even number. For example, four fans are set in each row, and correspondingly, four fans are set in each column; or six fans are set in each row, and correspondingly, six fans are set in each column; or eight fans are set in each row, and correspondingly, eight fans are set in each column.

[0041] According to some embodiments of the present invention, referring to Figure 1 The contraction section 12 is provided with a plurality of internal channels, and the number of internal channels is equal to and corresponds to the number of fans, that is, the contraction section 12 is provided with an independent internal channel for each DC wind speed, and the boundary curve of the internal channel is designed according to the Vidosinsky curve, and a honeycomb 121 is provided at the end of the contraction section 12 to effectively avoid the gap flow interference between fans, reduce the airflow boundary layer effect, and ensure the uniformity of the airflow. The honeycomb 121 is detachably connected to the wind tunnel, and the honeycomb 121 can be removed from the contraction section 12, and the honeycomb 121 can be selectively installed or not installed.

[0042] According to some embodiments of the present invention, referring to Figure 1 The multi-fan active control DC blowing micro wind tunnel 10 also includes: a computer system 5, a fan control system 3 includes: a pulse width modulation drive control board 32 and a single chip microcomputer 33, the pulse width modulation drive control board 32 is electrically connected to the single chip microcomputer 33, and the fan is connected to the single chip microcomputer 33. The computer system 5 controls the PWM duty cycle of each fan through the pulse width modulation drive control board 32 and the single chip microcomputer 33 to adjust the speed of each fan. For example, the PMW duty cycle can be 0%, 20%, 40%, 60%, 80%, 100%, etc.

[0043] The PMW duty cycle of each fan is uniformly managed by the computer system 5, and the PMW duty cycle of each fan is controlled by the combination of the pulse width modulation drive control board 32 and the single-chip microcomputer 33, thereby realizing dynamic adjustment of the fan speed and the wind speed profile of the experimental section 13, so as to realize intelligent adjustment.

[0044] For example, the fan is a 12V DC fan, and the fan control system 3 further includes a fan power supply 31, which is an ATX power supply. The ATX power supply is used to power the 12V DC fan, and the ATX power supply provides a stable 12V power supply to each fan.

[0045] According to some embodiments of the present invention, referring to Figure 1The multi-fan active controlled DC blowing micro wind tunnel 10 also includes: a computer system 5, a fan control system 3 is electrically connected to the computer system 5, a wind speed measurement system 4 includes: a three-axis mobile measuring frame 41 and a hot-wire anemometer 42, the hot-wire anemometer 42 includes a sensor probe 43, the three-axis mobile measuring frame 41 is installed in the experimental section 13, the hot-wire anemometer 42 is located outside the wind tunnel, the sensor probe 43 extends into the experimental section 13, and the sensor probe 43 is installed on the three-axis mobile measuring frame 41, and the computer system 5 is used to adjust the position of the three-axis mobile measuring frame 41.

[0046] The three-axis mobile measuring frame 41 can be adjusted in the three directions of X, Y and Z, and the position of the sensor probe 43 in the three directions of X, Y and Z can be adjusted, so that multi-point measurement can be achieved within the experimental section 13, and precise wind speed data can be collected. These wind speed data are transmitted to the computer system 5 through the hot wire anemometer 42 for data input of the wind speed profile debugging program.

[0047] According to the wind speed profile debugging method of the second embodiment of the present invention, which is applied to the multi-fan active control DC blowing micro wind tunnel 10 of the first embodiment of the present invention, the wind speed profile debugging method includes a convolutional neural network and a particle swarm optimization algorithm. The PWM duty cycle of the multi-fan array 2 is adjusted based on the convolutional neural network and the particle swarm optimization algorithm to achieve the target wind speed profile.

[0048] Convolutional neural network (CNN) can simulate complex velocity distribution through encoder / decoder structure due to its excellent performance in image recognition and data fitting. Particle swarm optimization (PSO) is a swarm intelligence optimization algorithm that can search for the optimal solution in multivariate space in an efficient way and has great potential for optimizing and controlling wind speed profiles.

[0049] By controlling the PWM duty cycle of the multi-fan array 2 to generate different speed combinations, the multi-fan array 2 includes a plurality of fans arranged in a determinant, and each fan has a different PWM duty cycle, such as 0%, 20%, 40%, 60%, 80%, 100%, thereby forming different wind speed modes. The convolutional neural network is used to output a two-dimensional array of wind speed profiles. In the process of achieving the target wind speed profile, the particle swarm optimization algorithm is used to automatically adjust the PWM duty cycle combination of the multi-fan array 2 to obtain the PWM duty cycle that best matches the preset wind speed profile, that is, to achieve the target wind speed profile.

[0050] According to the wind speed profile debugging method of the embodiment of the present invention, a convolutional neural network and a particle swarm optimization algorithm are used to accurately control the PWM duty cycle of the multi-fan array 2 to achieve the target wind speed profile.

[0051] According to some embodiments of the present invention, referring to Figure 1-Figure 2In the direction from the airflow inlet of the micro wind tunnel 1 to the airflow outlet of the micro wind tunnel 1, the micro wind tunnel 1 includes a power section 11, a contraction section 12, an experimental section 13 and a diffusion section 14 connected in sequence. The convolutional neural network is used to predict the wind speed profile of the multi-fan array 2 in the experimental section 13 at different duty ratios, and the duty ratio scheme of the multi-fan array 2 that best matches the target wind speed profile is determined according to the particle swarm optimization algorithm. By constructing a multi-fan array 2 and dynamically adjusting the speed of each fan, combined with the convolutional neural network and the particle swarm optimization algorithm, the precise control of the complex wind speed profile in the test section is achieved.

[0052] Reference Figure 2 , Figure 2 This is a schematic diagram of the structure of a convolutional neural network. The 8×8 rectangle below represents eight columns and eight rows of fans. Figure 2 The data in the rectangle on the left represents the duty cycle of each fan. Figure 2 The middle figure shows the structure of the convolutional neural network. Figure 2 The data in the rectangle on the right represents the normalized downstream average wind speed calculated by the convolutional neural network for the corresponding duty cycle of the fan on the left.

[0053] The wind speed profile debugging method uses a convolutional neural network as a prediction model and uses a particle swarm optimization algorithm to achieve precise control of the target wind speed profile. The optimization process includes the following steps:

[0054] (1) Target profile and initial population: Set the target wind speed profile shape in the experimental section 13, represented by a two-dimensional array. Initialize a set of particles, each particle represents a different combination of the fan PWM duty cycle, and the initial population covers all possible PWM duty cycle schemes.

[0055] (2) Fitness function calculation: The particle swarm optimization algorithm is used to predict the wind speed profile corresponding to each particle (i.e., PWM duty cycle scheme). The predicted result is compared with the target wind speed profile, and the fitness value is calculated by the mean square error. The smaller the fitness value, the higher the similarity between the PWM duty cycle scheme and the target wind speed profile.

[0056] (3) Update particle position and speed: Based on the update rules of the particle swarm optimization algorithm, gradually adjust the speed and position of the particle swarm so that it gradually converges to the global optimal solution. After each iteration, the particle solution with the best fitness is updated to the global optimal solution.

[0057] (4) Iteration termination condition: Repeat the fitness calculation and particle update until the termination condition is met (such as the number of iterations reaches the upper limit or the fitness value converges). The final PWM duty cycle combination is the optimal solution under the target wind speed profile.

[0058] According to some embodiments of the present invention, referring to Figure 1-Figure 2The convolutional neural network includes an encoder and a decoder. The encoder is used to receive a two-dimensional array of different PWM duty cycles of the multi-fan array 2, and the decoder generates a normalized two-dimensional array of downstream average wind speed on a cross-sectional lattice within the experimental section 13.

[0059] The convolutional neural network uses an encoder and a decoder. The input layer of the encoder receives the PWM duty cycle array of the multi-fan array 2, and the decoder outputs a two-dimensional array of wind speed profiles. Its design includes the following parts:

[0060] (1) The input layer receives a two-dimensional matrix consisting of a PWM duty cycle array, representing the speed setting of each fan. The encoder includes several convolutional layers, which extract the characteristics of the wind speed distribution under the fan configuration layer by layer; the depth of the encoding layer is adjusted according to the cross-sectional size of the experimental section 13 and the complexity of the wind speed pattern.

[0061] (2) The encoder feature output is reconstructed into a two-dimensional array of wind speed profiles by the decoder layer. The decoding layer uses deconvolution to enlarge the feature map layer by layer to achieve the mapping from the fan PMW duty cycle scheme to the wind speed profile.

[0062] (3) The loss function uses the mean square error (MSE) to calculate the error between the predicted wind speed profile and the target wind speed profile. The network weights are adjusted during the training process to minimize the error.

[0063] (4) The training data consists of multiple fan speed combinations and corresponding wind speed profiles. The data is trained using the gradient descent method. The Adam optimizer is used during training to improve the convergence speed, while batch normalization and L2 regularization are used to reduce the risk of overfitting.

[0064] According to some embodiments of the present invention, referring to Figure 1-Figure 2 The convolutional neural network is trained using training data, which consists of the measurement results of the average downstream wind speed on the cross-sectional lattice of the experimental section under different multi-fan array speed schemes. The measurement results are normalized based on the maximum downstream average wind speed on the cross-sectional lattice of the experimental section.

[0065] In order to ensure the accuracy of the wind speed profile debugging method, a large amount of training data under different fan speed schemes needs to be collected.

[0066] Specifically, the design of the training data scheme is divided into the following steps:

[0067] (1) Multi-fan array 2 speed scheme: By controlling the PWM duty cycle of each fan to generate different speed combinations, multiple representative PWM values ​​(e.g., 0%, 20%, 40%, 60%, 80%, 100%) are selected to form different wind speed modes, and the measurement cross-sectional array is arranged according to the actual size and requirements of the experimental section 13.

[0068] (2) Wind speed data measurement and normalization: Under each speed scheme, a measurement system consisting of a three-axis mobile measuring frame 41, a sensor probe 43 and a hot wire anemometer 42 is used to collect the downstream average wind speed of each point on the cross-sectional point array of the experimental section 13. In order to unify the wind speed scale, the measured wind speed data are normalized based on the maximum wind speed to form a two-dimensional array of different wind speed profiles.

[0069] The specific operation process of the wind speed profile debugging method is as follows:

[0070] (1) Fan speed scheme input: The user inputs the wind speed profile debugging requirements through the control interface and selects the wind speed profile shape within the experimental section 13.

[0071] (2) Automatic calculation of PWM duty cycle scheme: The program runs a convolutional neural network to predict the wind speed profile and executes a particle swarm optimization algorithm to automatically calculate the optimal PWM duty cycle combination to achieve the wind speed profile set by the user.

[0072] (3) Real-time feedback of wind speed profile: After the optimal PWM duty cycle scheme is generated, the PWM duty cycle signal of the multi-fan array 2 is automatically adjusted, and the wind speed profile is monitored in real time using the wind speed measurement system 4 in the experimental section 13 to confirm whether the preset target is achieved. If there is a deviation, the system can adjust the PWM wind speed profile debugging method M duty cycle scheme in real time based on the measurement results to ensure the accuracy of the wind speed profile.

[0073] The convolutional neural network adopts an encoder / decoder structure, in which the encoder is responsible for extracting the complex nonlinear relationship between the PWM duty cycle of the multi-fan array and the wind speed profile of the experimental section, while the decoder decodes the encoded information into a two-dimensional array of wind speed profiles.

[0074] The training data includes the actual measured data of wind speed profiles in experimental section 13 under various PWM duty cycle schemes. The data is normalized and input into the convolutional neural network for training to ensure that the model can accurately predict the wind speed distribution shape under different PWM duty cycle combinations. The output layer of the model generates a two-dimensional array of wind speed profiles to predict the distribution of downstream wind speed in experimental section 13.

[0075] In the process of achieving the target wind speed profile, the particle swarm optimization algorithm is used to automatically adjust the PWM duty cycle combination of the multi-fan array to obtain the PWM duty cycle scheme that best matches the target wind speed profile.

[0076] The particle swarm optimization algorithm initializes multiple particle groups, regards each particle as a wind speed control scheme, defines the fitness value according to the error between the wind speed profile predicted by the convolutional neural network and the target wind speed profile, and iterates and optimizes based on this fitness value. In each round of iteration, the particle swarm optimization algorithm dynamically adjusts the speed and position of the particles so that the particles gradually approach the optimal solution, thereby obtaining the PWM duty ratio configuration combination that is most similar to the target wind speed profile.

[0077] The wind speed profile debugging method is verified through actual wind tunnel experiments. A variety of different wind speed profile target values ​​are selected to record the PWM duty cycle scheme, the degree of conformity between the predicted and actual wind speed profiles, and the fitness changes of each adjustment. The results show that the wind speed profile debugging method of the present invention can achieve the preset wind speed profile in a relatively short time, and has high control accuracy and robustness.

[0078] The present invention significantly improves the controllability of the wind speed profile in the experimental section 13 of the micro wind tunnel 1 through the multi-fan active control and intelligent optimization control method, providing support for high-precision fluid mechanics experiments.

[0079] The wind speed profile debugging method of the present invention comprises the following steps:

[0080] First, the target wind speed profile is input by the user;

[0081] Then, the wind speed profile debugging program predicts the wind speed distribution under each PWM duty cycle combination through a convolutional neural network, and finds the optimal PWM duty cycle combination based on the particle swarm optimization algorithm to match the predicted wind speed profile with the target wind speed profile. The generated PWM duty cycle scheme will be implemented through the fan control system 3 to control the speed of the multi-fan array in real time.

[0082] In addition, the wind speed measurement system of the present invention can dynamically monitor the wind speed profile in the experimental section 13, and if necessary, can further fine-tune the PWM duty cycle scheme according to the feedback information to ensure that the profile meets expectations.

[0083] In summary, the multi-fan active control DC blowing micro wind tunnel and its wind speed profile debugging method of the present invention combines a multi-fan array, a convolutional neural network and a particle swarm optimization algorithm, which not only realizes the precise control of the wind speed profile, but also improves the automation and regulation efficiency of the system. This technology can meet the needs of complex flow field simulation in micro-scale wind tunnel experiments and provide an effective experimental means for related research and engineering applications.

[0084] In the description of this specification, the description with reference to the terms "some embodiments", "optionally", "further" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0085] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A multi-fan active controlled DC blowing micro wind tunnel (10), characterized in that: include: Breeze Tunnel (1); A multi-fan array (2), the multi-fan array (2) being arranged in the breeze tunnel (1), the multi-fan array (2) comprising a plurality of fans; A fan control system (3), the fan control system (3) being electrically connected to the multi-fan array (2) and used for controlling the rotation speeds of the multiple fans; A wind speed measurement system (4), wherein the wind speed measurement system (4) is used to obtain the wind speed in the micro wind tunnel (1).

2. The multi-fan active controlled DC blowing micro wind tunnel (10) according to claim 1, characterized in that: In the direction from the air flow inlet of the micro wind tunnel (1) to the air flow outlet of the micro wind tunnel (1), the micro wind tunnel (1) comprises a power section (11), a contraction section (12), an experimental section (13) and a diffusion section (14) which are connected in sequence, and the cross section of the micro wind tunnel (1) is square, the multi-fan array (2) is arranged on the power section (11), and the experimental section (13) is a light-permeable transparent member.

3. The multi-fan active controlled DC blowing micro wind tunnel (10) according to claim 2 is characterized in that: The number of the fans is an even square number, and the fans are arranged in a matrix and abut against each other in sequence.

4. The multi-fan active controlled DC blowing micro wind tunnel (10) according to claim 3 is characterized in that: The contraction section (12) is provided with a plurality of internal channels, the number of the internal channels is equal to and corresponds to the number of the fans, the boundary curves of the internal channels are designed according to the Vidosinski curve, and a honeycomb (121) is provided at the end of the contraction section (12), and the honeycomb (121) is detachably connected to the micro wind tunnel (1).

5. The multi-fan active controlled DC blowing micro wind tunnel (10) according to claim 3 is characterized in that: Also includes: A computer system (5), wherein the fan control system (3) comprises: a pulse width modulation drive control board (32) and a single chip microcomputer (33), wherein the pulse width modulation drive control board (32) is electrically connected to the single chip microcomputer (33), and the fan is connected to the single chip microcomputer (33), and the computer system (5) controls the PWM duty cycle of each of the fans through the pulse width modulation drive control board (32) and the single chip microcomputer (33) to adjust the rotation speed of each of the fans.

6. The multi-fan active controlled DC blowing micro wind tunnel (10) according to claim 2, characterized in that: Also includes: A computer system (5), the fan control system (3) is electrically connected to the computer system (5), the wind speed measurement system (4) comprises: a three-axis movable measuring frame (41) and a hot-wire anemometer (42), the hot-wire anemometer (42) comprises a sensor probe (43), the three-axis movable measuring frame (41) is installed in the experimental section (13), the hot-wire anemometer (42) is located outside the micro wind tunnel (1), the sensor probe (43) extends into the experimental section (13) and is installed on the three-axis movable measuring frame (41), and the computer system (5) is used to adjust the position of the three-axis movable measuring frame (41).

7. A wind speed profile debugging method, characterized in that: Applied to the multi-fan actively controlled DC blowing micro wind tunnel (10) according to any one of claims 1 to 6, the wind speed profile debugging method includes a convolutional neural network and a particle swarm optimization algorithm, and the PWM duty cycle of each of the fans is adjusted based on the convolutional neural network and the particle swarm optimization algorithm to achieve a target wind speed profile.

8. The wind speed profile debugging method according to claim 7, characterized in that: In the direction from the airflow inlet of the micro wind tunnel (1) to the airflow outlet of the micro wind tunnel (1), the micro wind tunnel (1) comprises a power section (11), a contraction section (12), a test section (13) and a diffusion section (14) which are connected in sequence. The convolutional neural network is used to predict the wind speed profile of the multi-fan array (2) in the experimental section (13) at different PWM duty cycles, and to determine the PWM duty cycle scheme of the multi-fan array (2) that best matches the target wind speed profile based on the particle swarm optimization algorithm.

9. The wind speed profile debugging method according to claim 8, characterized in that: The convolutional neural network comprises an encoder and a decoder, the encoder being used to receive a two-dimensional array of the multi-fan array (2) at different PWM duty cycles, and the decoder generating a two-dimensional array of normalized downstream average wind speed on a cross-sectional lattice within the experimental section (13).

10. The wind speed profile debugging method according to claim 9, characterized in that: The convolutional neural network is trained using training data, wherein the training data is composed of measurement results of the average downstream wind speed on the cross-sectional lattice of the experimental section (8) under different rotation speed schemes of the multi-fan array (2), and the measurement results are normalized based on the maximum average downstream wind speed on the cross-sectional lattice of the experimental section (8).

Images