Real-time adaptive adjustment system and method for wind tunnel boundary layer control

Abstract

The invention discloses a real-time adaptive adjustment system and method for wind tunnel boundary layer control, and relates to the technical field of wind tunnel boundary layer control. The data acquisition module acquires data at high frequency through multiple sensors and transmits the data through an ad hoc network, the data preprocessing module processes the data through specific filtering and dynamic normalization, and the boundary layer state evaluation module evaluates and predicts the boundary layer state by means of deep learning and a traditional model. The control decision module integrates hierarchical reinforcement learning and a model prediction control algorithm to determine a strategy; in the model verification and optimization step, the model is updated online through Bayesian fusion evaluation; in the man-machine interaction and remote monitoring step, remote monitoring and early warning are performed through interaction of virtual reality and augmented reality. According to the invention, wind tunnel experiment precision is improved, data are accurately collected and processed, and a boundary layer is accurately controlled; operation is convenient and safe, a visual interaction interface and remote monitoring are provided, a fault diagnosis technology is also provided, and development of related fields is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of wind tunnel boundary control, and in particular to a real-time adaptive adjustment system and method for wind tunnel boundary layer control. Background Art

[0002] Wind tunnel experiments are an important tool for studying aerodynamic properties in a wide range of fields, including aerospace, automotive manufacturing, and construction engineering. The wind tunnel boundary layer, a crucial component of the wind tunnel flow field, directly impacts the accuracy and reliability of experimental results. However, current wind tunnel boundary layer control faces numerous challenges.

[0003] During wind tunnel operation, the internal flow field is influenced by numerous factors, including the shape and size of the experimental model, the structure of the wind tunnel itself, and changes in external environmental conditions. These factors lead to complex and variable boundary layer states, making them difficult to stably control. Traditional boundary layer control methods, mostly based on fixed parameters and preset rules, lack the ability to effectively respond to real-time changes in the boundary layer. When the boundary layer becomes unstable or deviates from its ideal state, traditional methods are unable to adjust in a timely manner, resulting in large errors in experimental results and failing to meet the requirements of high-precision experiments.

[0004] From the perspective of technical implementation, existing boundary layer control technologies have obvious deficiencies in data acquisition, processing, and regulation execution. In terms of data acquisition, the accuracy and layout of sensors make it difficult to fully and accurately obtain various parameters of the boundary layer, resulting in inaccurate assessments of the boundary layer state. Data processing technology is also relatively simple, unable to fully tap the potential information in the data, and difficult to make accurate predictions about the changing trends of the boundary layer. In the regulation execution link, the response speed, regulation accuracy, and coordination of the actuators are poor, making it impossible to implement control strategies quickly and accurately, and difficult to achieve effective regulation of the boundary layer. In addition, the structures and experimental requirements of different wind tunnels vary greatly, and the existing control technology lacks versatility and is difficult to be widely used in various wind tunnel scenarios. These problems have seriously restricted the development of wind tunnel experimental technology and hindered the in-depth advancement of aerodynamic research in related fields. An innovative wind tunnel boundary layer control technology is urgently needed to solve these problems. Summary of the Invention

[0005] The present invention proposes a real-time adaptive adjustment system and method for wind tunnel boundary layer control to solve the problems mentioned in the above-mentioned prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a real-time adaptive adjustment system for wind tunnel boundary layer control, comprising: Data acquisition module: Pressure sensors, velocity sensors, temperature sensors, humidity sensors, and wall shear stress sensors are installed inside the wind tunnel, in the boundary layer area, and in the external environment of the wind tunnel. The sensors collect data, and the sensor nodes use distributed self-organizing network technology to automatically optimize the data transmission path and transmit it to the data processing center. Data preprocessing module: The collected sensor data is filtered using a filtering algorithm based on wavelet decomposition and adaptive threshold fusion. The threshold is adaptively determined according to the statistical characteristics of the data to remove noise interference. The filtered data is then normalized using a dynamic normalization method to dynamically adjust the normalization interval according to the real-time distribution range of the data. The formula is: Adapting to changes in boundary layer conditions, where and It is the minimum and maximum value of data that changes dynamically over time; Boundary layer state assessment module: Introduces a boundary layer state prediction model based on deep learning, combines preprocessed data and historical data, and predicts the future state of the boundary layer. The model is based on convolutional neural network (CNN) and long short-term memory network (LSTM), and combines traditional boundary layer theoretical models to comprehensively calculate the boundary layer thickness. , displacement thickness , momentum thickness θ and boundary layer transition position Parameters; the boundary layer transition position prediction formula is ,in is the local Reynolds number, Tu is the turbulence intensity, and f is the nonlinear function obtained through deep learning model training; Control Decision Module: This module uses an algorithm based on the fusion of hierarchical reinforcement learning and model predictive control (MPC) to determine the regulation strategy. The control task is divided into two levels: high-level decision-making and low-level execution. High-level decision-making is based on the reinforcement learning algorithm, which determines the regulation direction and target based on the state and target of the boundary layer. Low-level execution uses the MPC algorithm. Based on the high-level decision-making, it predicts state changes over a period of time in the future and optimizes the regulation action sequence based on the real-time state and dynamic model of the boundary layer. An adaptive exploration-exploitation strategy is also introduced to dynamically adjust the exploration and exploitation ratio based on the state of the boundary layer. Adjustment execution module: The adjustment actuator is driven by a micro-electromechanical system (MEMS). Micro-nano processing technology is used to manufacture a wall structure with continuously adjustable roughness. The shape and height of the wall microstructure are controlled by external excitation of electric and magnetic fields to achieve continuous adjustment of the wall roughness. The control signal of the actuator adopts coded modulation technology to achieve anti-interference capability of signal transmission. Model verification and optimization module: During the wind tunnel experiment, the boundary layer state data before and after adjustment are compared with the results of the theoretical model and the prediction model in real time. The Bayesian model averaging (BMA) method is used to fuse and evaluate the prediction results of different models. The weights and parameters of the model are dynamically adjusted according to the fusion results. The formula is: ,in is the prediction result under data D, is the i-th model, is the weight of the ith model, and ; Human-computer interaction and remote monitoring module: Develop a human-computer interaction interface based on virtual reality and augmented reality technologies. Operators use the equipment to observe the boundary layer flow state and adjustment process inside the wind tunnel, and obtain parameters and adjustment instructions in real time in the actual wind tunnel scene through the equipment. At the same time, the system supports remote monitoring and control functions, and transmits wind tunnel operation data and status information to the remote monitoring center in real time via wireless networks. Operators can remotely formulate adjustment strategies and control the execution mechanism.

[0007] Furthermore, it also includes: Environmental perception and compensation module: Real-time monitoring of the wind tunnel's external atmospheric pressure, ambient temperature, wind speed and direction. Based on changes in environmental parameters, a fuzzy control algorithm is used to compensate and adjust the boundary layer control strategy. The fuzzy control rules are established based on expert experience and experimental data. According to different environmental parameter combinations and boundary layer states, the action amplitude and frequency of the regulating actuator are automatically adjusted.

[0008] Furthermore, the data acquisition module adopts sensor fault diagnosis and fault tolerance technology to determine in real time whether a sensor fault occurs through integrated analysis and residual detection of sensor data; when a sensor fault is detected, a state estimation method based on Kalman filtering is used to reconstruct the data of the faulty sensor, and the sampling strategy and data fusion algorithm of each sensor are automatically adjusted.

[0009] Furthermore, the control decision module introduces a multi-objective optimization mechanism. In addition to optimizing the stability and uniformity of the boundary layer, it also considers the energy consumption of the adjustment process, the adjustment time, and the impact on the wind tunnel test parameters. The non-dominated sorting genetic algorithm (NSGA-II) is used to optimize and solve the objectives, obtaining a set of Pareto optimal solutions, from which the operator selects the adjustment strategy according to actual needs.

[0010] Furthermore, the regulation execution module adopts a collaborative control strategy. For different types of actuators, a coupling dynamic model is established to analyze the interactions and influences between them. A distributed collaborative control algorithm is used to enable different actuators to work together and avoid mutual interference and conflict during the regulation process.

[0011] Furthermore, the model verification and optimization module also adopts a real-time data-driven model updating method; using an online learning algorithm, the theoretical model and the prediction model are updated and optimized according to the boundary layer data collected in real time.

[0012] Furthermore, the human-computer interaction and remote monitoring module is also equipped with early warning and decision-making assistance functions. The system automatically issues early warning information and provides response suggestions and decision-making plans when abnormal conditions or dangerous conditions are detected through real-time analysis and prediction of boundary layer status data. At the same time, knowledge graph technology is used to manage and mine wind tunnel test data and adjustment experience to provide decision support for operators.

[0013] A method for applying the real-time adaptive adjustment algorithm system for wind tunnel boundary layer control includes: Data collection steps: Install sensors at the wind tunnel location to collect boundary layer and ambient pressure, velocity, temperature, humidity, and wall shear stress data at a high-frequency sampling rate, and transmit the data via distributed ad hoc networking technology; Data preprocessing steps: Use a filtering algorithm based on wavelet decomposition and adaptive threshold fusion to reduce noise on the collected data, and then use a dynamic normalization method to normalize the data according to its real-time distribution range; Boundary layer state assessment steps: Utilize a deep learning-based prediction model combined with traditional theoretical models to calculate boundary layer thickness, displacement thickness, momentum thickness, and transition position parameters, and predict the future state of the boundary layer; Control decision-making steps: Adopting a fusion algorithm of hierarchical reinforcement learning and model predictive control, combined with an adaptive exploration-exploitation strategy, to determine the regulation strategy while considering multi-objective optimization; Adjustment execution steps: The adjustment action is driven by the micro-electromechanical system, the control signal is transmitted using coded modulation technology, and the actuators work together; Model validation and optimization steps: real-time comparison of pre- and post-adjustment data with model results, use of Bayesian model averaging to fuse and evaluate different models, and use of online learning algorithms to update the model; Human-computer interaction and remote monitoring steps: Human-computer interaction is achieved through virtual reality and augmented reality technologies, remote monitoring and control are supported, and the system is equipped with early warning and decision-making assistance functions.

[0014] Furthermore, it also includes: Environmental perception and compensation steps: Real-time monitoring of the external meteorological conditions of the wind tunnel, and the use of fuzzy control algorithms to compensate and adjust the boundary layer control strategy according to changes in environmental parameters.

[0015] Sensor fault diagnosis and fault tolerance steps: integrate and analyze sensor data and perform residual detection to determine sensor faults, use Kalman filtering to reconstruct faulty sensor data, and adjust the sampling and fusion strategies of each sensor.

[0016] Compared with the existing technology, the beneficial effects of the present invention are: To enhance experimental accuracy, the dense layout of high-precision, multi-type sensors and high-frequency sampling enable comprehensive and accurate acquisition of boundary layer parameters. Advanced data processing algorithms effectively remove noise interference and dynamically normalize data, enabling more precise assessment of boundary layer conditions. A predictive model combining deep learning with traditional theory accurately calculates key parameters and predicts future conditions, providing strong support for precise control. The high response speed and precise adjustment capabilities of the regulating actuator ensure that the boundary layer quickly approaches the ideal state, significantly improving the accuracy of experimental results and providing researchers with more reliable data.

[0017] From the perspective of adaptability and flexibility, the system demonstrates exceptional real-time adaptive regulation capabilities. Whether responding to changes in experimental conditions within the wind tunnel or interference from external environmental factors, the system's environmental perception and compensation module and adaptive control algorithm enable timely adjustments to control strategies. A multi-objective optimization mechanism enables flexible selection of adjustment strategies based on diverse experimental requirements. Furthermore, the system utilizes advanced technologies such as distributed ad hoc networking and intelligent material actuation, enhancing its versatility and enabling stable operation in a variety of wind tunnel types.

[0018] This patented achievement also offers significant advantages in operational convenience and safety. The human-computer interaction interface, based on virtual reality and augmented reality, allows operators to intuitively and immersively observe and control the experimental process. Remote monitoring and intelligent early warning functions allow operators to monitor the status of the wind tunnel at all times, address abnormalities promptly, and ensure experimental safety. Furthermore, the system's fault diagnosis and fault-tolerance technology, model verification, and optimization mechanisms enhance system reliability and stability, reduce equipment failures and maintenance costs, and improve the overall efficiency of wind tunnel experiments, effectively promoting the development of aerodynamic research in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic block diagram of the real-time adaptive adjustment system for wind tunnel boundary layer control proposed by the present invention; Figure 2 A schematic block diagram of the steps of the real-time adaptive adjustment algorithm for wind tunnel boundary layer control proposed by the present invention; Figure 3 A schematic diagram showing a comparison of the control accuracy of boundary layer thickness of different systems for real-time adaptive adjustment of wind tunnel boundary layer control proposed by the present invention; Figure 4This is a schematic diagram showing how the accuracy of experimental results of different systems changes over time in the real-time adaptive adjustment of the wind tunnel boundary layer control proposed in the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations 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 should not be understood as limiting the present invention.

[0022] 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. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, 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 it can be indirectly connected through an intermediate medium, or it can be a connection between the 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. The present invention will be further described in detail below with reference to the accompanying drawings.

[0023] Reference Figures 1 to 4 The present invention proposes a specific embodiment of a real-time adaptive adjustment system and method for wind tunnel boundary layer control, including: Data acquisition module: Pressure sensors, velocity sensors, temperature sensors, humidity sensors, and wall shear stress sensors are installed inside the wind tunnel, in the boundary layer area, and in the external environment of the wind tunnel. The sensors collect data, and the sensor nodes use distributed self-organizing network technology to automatically optimize the data transmission path and transmit it to the data processing center. Data preprocessing module: The collected sensor data is filtered using a filtering algorithm based on wavelet decomposition and adaptive threshold fusion. The threshold is adaptively determined according to the statistical characteristics of the data to remove noise interference. The filtered data is then normalized using a dynamic normalization method to dynamically adjust the normalization interval according to the real-time distribution range of the data. The formula is: Adapting to changes in boundary layer conditions, where and It is the minimum and maximum value of data that changes dynamically over time; Boundary layer state assessment module: Introduces a boundary layer state prediction model based on deep learning, combines preprocessed data and historical data, and predicts the future state of the boundary layer. The model is based on convolutional neural network (CNN) and long short-term memory network (LSTM), and combines traditional boundary layer theoretical models to comprehensively calculate the boundary layer thickness. , displacement thickness , momentum thickness θ and boundary layer transition position Parameters; the boundary layer transition position prediction formula is ,in is the local Reynolds number, Tu is the turbulence intensity, and f is the nonlinear function obtained through deep learning model training; Control Decision Module: This module uses an algorithm based on the fusion of hierarchical reinforcement learning and model predictive control (MPC) to determine the regulation strategy. The control task is divided into two levels: high-level decision-making and low-level execution. High-level decision-making is based on the reinforcement learning algorithm, which determines the regulation direction and target based on the state and target of the boundary layer. Low-level execution uses the MPC algorithm. Based on the high-level decision-making, it predicts state changes over a period of time in the future and optimizes the regulation action sequence based on the real-time state and dynamic model of the boundary layer. An adaptive exploration-exploitation strategy is also introduced to dynamically adjust the exploration and exploitation ratio based on the state of the boundary layer. Adjustment execution module: The adjustment actuator is driven by a micro-electromechanical system (MEMS). Micro-nano processing technology is used to manufacture a wall structure with continuously adjustable roughness. The shape and height of the wall microstructure are controlled by external excitation of electric and magnetic fields to achieve continuous adjustment of the wall roughness. The control signal of the actuator adopts coded modulation technology to achieve anti-interference capability of signal transmission. Model verification and optimization module: During the wind tunnel experiment, the boundary layer state data before and after adjustment are compared with the results of the theoretical model and the prediction model in real time. The Bayesian model averaging (BMA) method is used to fuse and evaluate the prediction results of different models. The weights and parameters of the model are dynamically adjusted according to the fusion results. The formula is: ,in is the prediction result under data D, is the i-th model, is the weight of the ith model, and ; Human-computer interaction and remote monitoring module: Develop a human-computer interaction interface based on virtual reality and augmented reality technologies. Operators use the equipment to observe the boundary layer flow state and adjustment process inside the wind tunnel, and obtain parameters and adjustment instructions in real time in the actual wind tunnel scene through the equipment. At the same time, the system supports remote monitoring and control functions, and transmits wind tunnel operation data and status information to the remote monitoring center in real time via wireless networks. Operators can remotely formulate adjustment strategies and control the execution mechanism.

[0024] The present invention also includes: Environmental perception and compensation module: Real-time monitoring of the wind tunnel's external atmospheric pressure, ambient temperature, wind speed and direction. Based on changes in environmental parameters, a fuzzy control algorithm is used to compensate and adjust the boundary layer control strategy. The fuzzy control rules are established based on expert experience and experimental data. According to different environmental parameter combinations and boundary layer states, the action amplitude and frequency of the regulating actuator are automatically adjusted.

[0025] In the present invention, the data acquisition module adopts sensor fault diagnosis and fault tolerance technology to determine in real time whether a sensor fault occurs through integrated analysis and residual detection of sensor data; when a sensor fault is detected, a state estimation method based on Kalman filtering is used to reconstruct the data of the faulty sensor, and the sampling strategy and data fusion algorithm of each sensor are automatically adjusted.

[0026] In the present invention, the control decision module introduces a multi-objective optimization mechanism. In addition to optimizing the stability and uniformity of the boundary layer, it also considers the energy consumption of the adjustment process, the adjustment time, and the impact on the wind tunnel test parameters. The non-dominated sorting genetic algorithm (NSGA-II) is used to optimize and solve the objectives, obtaining a set of Pareto optimal solutions, from which the operator selects the adjustment strategy according to actual needs.

[0027] In the present invention, the regulation execution module adopts a collaborative control strategy. For different types of actuators, a coupling dynamic model is established to analyze the interactions and influences between them. A distributed collaborative control algorithm is used to achieve collaborative work of different actuators to avoid mutual interference and conflict during the regulation process.

[0028] In the present invention, the model verification and optimization module also adopts a real-time data-driven model updating method; using an online learning algorithm, the theoretical model and the prediction model are updated and optimized according to the boundary layer data collected in real time.

[0029] In the present invention, the human-computer interaction and remote monitoring module is also equipped with early warning and decision-making assistance functions. The system automatically issues early warning information and provides response suggestions and decision-making plans when abnormal conditions or dangerous conditions are detected through real-time analysis and prediction of boundary layer status data. At the same time, knowledge graph technology is used to manage and mine wind tunnel test data and adjustment experience to provide decision support for operators.

[0030] The present invention also discloses a real-time adaptive adjustment method using the wind tunnel boundary layer control, comprising: Data collection steps: Install sensors at the wind tunnel location to collect boundary layer and ambient pressure, velocity, temperature, humidity, and wall shear stress data at a high-frequency sampling rate, and transmit the data via distributed ad hoc networking technology; Data preprocessing steps: Use a filtering algorithm based on wavelet decomposition and adaptive threshold fusion to reduce noise on the collected data, and then use a dynamic normalization method to normalize the data according to its real-time distribution range; Boundary layer state assessment steps: Utilize a deep learning-based prediction model combined with traditional theoretical models to calculate boundary layer thickness, displacement thickness, momentum thickness, and transition position parameters, and predict the future state of the boundary layer; Control decision-making steps: Adopting a fusion algorithm of hierarchical reinforcement learning and model predictive control, combined with an adaptive exploration-exploitation strategy, to determine the regulation strategy while considering multi-objective optimization; Adjustment execution steps: The adjustment action is driven by the micro-electromechanical system, the control signal is transmitted using coded modulation technology, and the actuators work together; Model validation and optimization steps: real-time comparison of pre- and post-adjustment data with model results, use of Bayesian model averaging to fuse and evaluate different models, and use of online learning algorithms to update the model; Human-computer interaction and remote monitoring steps: Human-computer interaction is achieved through virtual reality and augmented reality technologies, remote monitoring and control are supported, and the system is equipped with early warning and decision-making assistance functions.

[0031] The present invention also includes: Environmental perception and compensation steps: Real-time monitoring of the external meteorological conditions of the wind tunnel, and the use of fuzzy control algorithms to compensate and adjust the boundary layer control strategy according to changes in environmental parameters.

[0032] Sensor fault diagnosis and fault tolerance steps: integrate and analyze sensor data and perform residual detection to determine sensor faults, use Kalman filtering to reconstruct faulty sensor data, and adjust the sampling and fusion strategies of each sensor.

[0033] Example 1

[0034] 1. System Hardware Deployment

[0035] A comprehensive sensor network was constructed within the wind tunnel, within the boundary layer, and at relevant locations in the external environment. High-precision pressure sensors were installed at various heights and along the wind tunnel at intervals required by technical specifications to measure pressure distribution within the boundary layer. Nearby, laser Doppler velocimeters with an accuracy of ±0.02 m / s were deployed as velocity sensors, ensuring accurate airflow velocity information. Temperature sensors (accuracy of ±0.1 K), humidity sensors (accuracy of ±1% RH), and wall shear stress sensors (accuracy of ±0.01 Pa) were also spaced at appropriate intervals to comprehensively collect various data. These sensors were connected to a data acquisition card, which collected sensor data at a frequency of 1000-5000 Hz and transmitted it to a data processing center via distributed ad hoc networking technology. This ad hoc network employed the Ad-Hoc On-Demand Distance Vector (AODV) routing protocol to ensure reliable and real-time data transmission.

[0036] 2. System Software Process

[0037] (I) Data preprocessing In the data processing center, after receiving the data collected by the sensor, the first step is to use a filtering algorithm based on multi-scale wavelet decomposition and adaptive threshold fusion to perform noise reduction. Taking wind speed data as an example, it is decomposed into sub-signals of different frequencies by multi-scale wavelet decomposition. For each scale detail coefficient, according to its local standard deviation and the preset threshold coefficient k (such as k=3), and adaptively determine the threshold The detail coefficients are processed by the threshold to remove the noise components, and then wavelet reconstruction is performed to obtain the filtered wind speed data.

[0038] Next, the filtered data is dynamically normalized. Taking the pressure data P as an example, the minimum value of the pressure data in the current time period is obtained in real time. and maximum value , according to the formula Normalization is performed to map the pressure data to the [0,1] interval for subsequent analysis.

[0039] (2) Boundary layer state assessment

[0040] The boundary layer state is assessed by combining a deep learning-based boundary layer state prediction model with a traditional boundary layer theory model. The deep learning model uses a convolutional neural network (CNN) and a long short-term memory (LSTM) network to construct the prediction model. Preprocessed sensor data, such as pressure, velocity, and temperature, are arranged according to time series and spatial location and serve as input to the CNN. The CNN extracts spatial features from the data through convolutional and pooling layers, and then passes its output to the LSTM. The LSTM processes the time series information and learns how the boundary layer state changes over time.

[0041] At the same time, combined with the traditional boundary layer theoretical model, the boundary layer thickness is calculated , displacement thickness , momentum thickness θ and boundary layer transition position Taking the boundary layer thickness calculation as an example, according to the formula , where the kinematic viscosity coefficient is It can be calculated by empirical formula based on the current temperature and gas properties. The distance x from the wind tunnel entrance to the measurement point is determined by the sensor position, and the incoming flow velocity The boundary layer transition position is predicted based on the formula , where the local Reynolds number The turbulence index Tu is calculated from velocity fluctuations measured by a velocity sensor, and f is a nonlinear function trained using a deep learning model. By combining the deep learning model's predictions with those calculated using traditional theoretical models, a comprehensive assessment of boundary layer stability, turbulence intensity, and other state information is provided.

[0042] (3) Control decision-making

[0043] The control strategy is determined using a fusion algorithm of hierarchical reinforcement learning and model predictive control (MPC). The control task is divided into two levels: high-level decision-making and low-level execution. In the high-level decision-making, the state space is defined as the key parameters of the boundary layer (such as boundary layer thickness, displacement thickness, turbulence intensity, etc.) and the current state of the control actuator; the action space is the adjustment direction and approximate amplitude of the control actuator. The reward function is determined based on factors such as the degree to which the boundary layer state approaches the ideal state and the energy consumption of the control. The Deep Q-Network (DQN) algorithm is used for training to learn the optimal control strategy.

[0044] At the low-level, the MPC algorithm predicts the boundary layer's state over a period of time based on its real-time state and dynamic model (e.g., a simplified discrete model of the Navier-Stokes equations). By optimizing the objective function (e.g., minimizing the deviation between the boundary layer state and the ideal state and the magnitude of the regulatory action), a sequence of regulatory actions is generated. This optimization process takes into account system constraints, such as the actuator's adjustment range and response speed limits.

[0045] At the same time, an adaptive exploration-exploitation strategy is introduced to dynamically adjust the ratio of exploration to exploitation based on the uncertainty of the boundary layer state (such as the error of the prediction model). When uncertainty is high, the exploration ratio is increased, and more different adjustment actions are tried; when uncertainty is low, the ratio of exploitation of the learned optimal strategy is increased.

[0046] (IV) Regulation and implementation

[0047] The regulating actuator uses a shape memory alloy (SMA)-driven microelectromechanical system (MEMS) to control wall suction / blow and adjust wall roughness. For the wall suction / blow mechanism, an SMA microactuator is connected to the valve in the suction / blow duct. Upon receiving a control signal, the SMA microactuator deforms according to the current, precisely controlling the valve opening and adjusting the suction / blow rate. Its response speed can reach milliseconds, enabling rapid response to changes in boundary layer conditions.

[0048] For wall roughness adjustment, a continuously adjustable wall structure is fabricated using micro-nanofabrication technology. By applying an electric or magnetic field to the wall microstructure, the shape and height of the microstructure are altered, enabling continuous and precise adjustment of the wall roughness. The control signal is transmitted using FSK (frequency shift keying) coded modulation technology to improve signal interference resistance. During the adjustment process, different actuators (such as the suction / blowing device and the wall roughness adjustment device) work together through a collaborative control strategy. By establishing a coupled dynamics model and analyzing their interactions and influences, a distributed collaborative control algorithm is implemented to ensure that the adjustment actions of each actuator are coordinated, avoiding mutual interference and improving adjustment effectiveness and efficiency.

[0049] (V) Model verification and optimization

[0050] During the wind tunnel experiment, the boundary layer state data before and after adjustment are compared with the results of the theoretical model and the prediction model in real time. The Bayesian model averaging (BMA) method is used to integrate and evaluate the prediction results of different models. Assume that there are M different models , for each model, calculate its weight based on historical data and current experimental data The historical data is the record of boundary layer state and related parameters accumulated in the early stage of wind tunnel experiment, covering boundary layer data under different working conditions and adjustment strategies, which constitutes the basic basis for training and evaluation of each model. When the current experiment is carried out, the boundary layer state data (such as velocity profile, turbulence intensity, etc.) is collected in real time, and combined with the historical data, each model is trained. Evaluation: By analyzing the model's prediction accuracy for similar boundary layer scenarios in historical data and its fit with actual measurements under current experimental data, the likelihood of the model is calculated. The formula is ,in is the prediction result under data D, is the prediction result of the i-th model under data D. Based on the fusion results, the model weights and parameters are dynamically adjusted. Using online learning algorithms, such as online support vector machines (OnlineSVM), the theoretical and prediction models are continuously updated and optimized based on real-time boundary layer data. Online learning algorithms can quickly adapt to changes in boundary layer conditions, improving the real-time performance and adaptability of the model.

[0051] (6) Human-computer interaction and remote monitoring

[0052] Develop a human-computer interaction interface based on virtual reality (VR) and augmented reality (AR) technologies. In a VR environment, operators can immersively observe the boundary layer flow state inside the wind tunnel and use handles to view details of the boundary layer from different angles, such as velocity vector distribution and pressure cloud maps. In an AR environment, operators at the wind tunnel site can use smart terminal devices such as tablets or smart glasses to see key parameters superimposed on the actual wind tunnel scene in real time, such as boundary layer thickness and the current status of the adjustment actuator. At the same time, the system supports remote monitoring and control functions, transmitting wind tunnel operation data and status information to a remote monitoring center in real time via a 5G high-speed wireless network. In the remote monitoring center, operators can use computers or mobile devices, with the help of VR or AR technology, to remotely formulate adjustment strategies and control actuators, just like on-site operations.

[0053] The system also features intelligent early warning and decision-making support. Through real-time analysis and prediction of boundary layer status data, the system automatically issues warning information, such as audible alarms and pop-up prompts, when it detects that the boundary layer status may exceed the safe range or an abnormal situation occurs. At the same time, based on preset rules and historical experience, the system provides corresponding response suggestions and decision-making plans to help operators respond quickly. Knowledge graph technology is used to manage and mine wind tunnel test data and adjustment experience, providing intelligent decision-making support for operators. For example, when encountering a specific boundary layer anomaly, the system can quickly search the knowledge graph to find experience and successful cases in handling similar situations, providing reference for operators.

[0054] 3. Data Representation and Interpretation

[0055] The data in the table clearly demonstrates that the proposed system offers significant improvements over conventional systems in multiple areas. The significantly improved accuracy of boundary layer thickness control ensures that boundary layer conditions during wind tunnel experiments are more stably maintained within the ideal range, reducing experimental errors caused by boundary layer instability. The 10-15 percentage point increase in experimental result accuracy means that researchers can obtain more reliable data, providing a more solid foundation for research in related fields. The system's response time has been shortened by 50%-67%, enabling faster response to changes in boundary layer conditions and timely adjustment of control strategies, further improving experimental stability and reliability. Energy savings have increased by 20-30 percentage points compared to conventional systems, demonstrating the proposed system's energy-saving advantages and reducing wind tunnel operating costs. The operational ease of use score has increased from 2-3 to 4-5, an increase of approximately 67%-150%. The human-computer interaction interface and remote monitoring capabilities based on VR and AR technologies enable more intuitive and convenient operation and monitoring for operators, significantly improving work efficiency and reducing the possibility of operational errors. These improvements, taken together, have comprehensively enhanced the quality and efficiency of wind tunnel experiments, providing strong support for the development of related fields.

[0056] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A real-time adaptive adjustment system for wind tunnel boundary layer control, characterized in that: include: Data acquisition module: Pressure sensors, velocity sensors, temperature sensors, humidity sensors, and wall shear stress sensors are installed inside the wind tunnel, in the boundary layer area, and in the external environment of the wind tunnel. The sensors collect data, and the sensor nodes use distributed self-organizing network technology to automatically optimize the data transmission path and transmit it to the data processing center. Data preprocessing module: The collected sensor data is filtered using a filtering algorithm based on wavelet decomposition and adaptive threshold fusion. The threshold is adaptively determined according to the statistical characteristics of the data to remove noise interference. The filtered data is then normalized using a dynamic normalization method to dynamically adjust the normalization interval according to the real-time distribution range of the data. The formula is: Adapting to changes in boundary layer conditions, where and It is the minimum and maximum value of data that changes dynamically over time; Boundary layer state assessment module: Introduces a boundary layer state prediction model based on deep learning, combines preprocessed data and historical data, and predicts the future state of the boundary layer. The model is based on convolutional neural network (CNN) and long short-term memory network (LSTM), and combines traditional boundary layer theoretical models to comprehensively calculate the boundary layer thickness. , displacement thickness , momentum thickness θ and boundary layer transition position Parameters; the boundary layer transition position prediction formula is ,in is the local Reynolds number, Tu is the turbulence intensity, and f is the nonlinear function obtained through deep learning model training; Adjustment execution module: The adjustment actuator is driven by a micro-electromechanical system (MEMS), and micro-nano processing technology is used to manufacture a wall structure with continuously adjustable roughness. The shape and height of the wall microstructure are controlled by external excitation of electric and magnetic fields to achieve continuous adjustment of the wall roughness. The control signal of the actuator adopts coded modulation technology to achieve anti-interference capability of signal transmission.

2. The real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 1, characterized in that: Also includes: Control Decision Module: This module uses an algorithm based on the fusion of hierarchical reinforcement learning and model predictive control (MPC) to determine the regulation strategy. The control task is divided into two levels: high-level decision-making and low-level execution. High-level decision-making is based on the reinforcement learning algorithm, which determines the regulation direction and target based on the state and target of the boundary layer. Low-level execution uses the MPC algorithm. Based on the high-level decision-making, it predicts state changes over a period of time in the future and optimizes the regulation action sequence based on the real-time state and dynamic model of the boundary layer. An adaptive exploration-exploitation strategy is also introduced to dynamically adjust the exploration and exploitation ratio based on the state of the boundary layer. Model verification and optimization module: During the wind tunnel experiment, the boundary layer state data before and after adjustment are compared with the results of the theoretical model and the prediction model in real time. The Bayesian model averaging (BMA) method is used to fuse and evaluate the prediction results of different models. The weights and parameters of the model are dynamically adjusted according to the fusion results. The formula is: ,in is the prediction result under data D, is the i-th model, is the weight of the ith model, and ; Human-computer interaction and remote monitoring module: Develop a human-computer interaction interface based on virtual reality and augmented reality technologies. Operators use the device to observe the boundary layer flow state and adjustment process inside the wind tunnel, and obtain parameters and adjustment instructions in real time through the device in the actual wind tunnel scene. The system also supports remote monitoring and control functions, transmitting wind tunnel operation data and status information in real time to the remote monitoring center via wireless network, allowing operators to remotely formulate adjustment strategies and control the execution mechanism. Environmental perception and compensation module: Real-time monitoring of the wind tunnel's external atmospheric pressure, ambient temperature, wind speed and direction. Based on changes in environmental parameters, a fuzzy control algorithm is used to compensate and adjust the boundary layer control strategy. The fuzzy control rules are established based on expert experience and experimental data. According to different environmental parameter combinations and boundary layer states, the action amplitude and frequency of the regulating actuator are automatically adjusted.

3. The real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 1, characterized in that: The data acquisition module adopts sensor fault diagnosis and fault tolerance technology to determine whether the sensor is faulty in real time through integrated analysis of sensor data and residual detection. When a sensor fault is detected, the state estimation method based on Kalman filtering is used to reconstruct the data of the faulty sensor, and the sampling strategy and data fusion algorithm of each sensor are automatically adjusted.

4. The real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 2, characterized in that: The control decision module introduces an optimization mechanism that, in addition to optimizing the stability and uniformity of the boundary layer, also considers the energy consumption of the adjustment process, the adjustment time, and the impact on the wind tunnel test parameters. The non-dominated sorting genetic algorithm (NSGA-II) is used to optimize and solve the objectives, obtaining a set of Pareto optimal solutions. The operator then selects the adjustment strategy based on actual needs.

5. The real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 1, characterized in that: The regulation execution module adopts a collaborative control strategy. For different types of actuators, a coupling dynamic model is established to analyze the interactions and influences between them. A distributed collaborative control algorithm is used to enable different actuators to work together to avoid mutual interference during the regulation process.

6. The real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 2, characterized in that: The model validation and optimization module also adopts a real-time data-driven model updating method; using an online learning algorithm, the theoretical model and the prediction model are updated and optimized based on the boundary layer data collected in real time.

7. The real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 2, characterized in that: The human-computer interaction and remote monitoring module is also equipped with early warning and decision-making assistance functions. The system automatically issues early warning information and provides response suggestions and decision-making plans when abnormal conditions or dangerous conditions are detected through real-time analysis and prediction of boundary layer status data. It also uses knowledge graph technology to manage and mine wind tunnel test data and adjustment experience to provide decision support for operators.

8. A method for applying the real-time adaptive adjustment system for wind tunnel boundary layer control according to any one of claims 1 to 7, characterized in that: include: Data collection steps: Install sensors at the wind tunnel location to collect boundary layer and ambient pressure, velocity, temperature, humidity, and wall shear stress data at a high-frequency sampling rate, and transmit the data via distributed ad hoc networking technology; Data preprocessing steps: Use a filtering algorithm based on wavelet decomposition and adaptive threshold fusion to reduce noise on the collected data, and then use a dynamic normalization method to normalize the data according to its real-time distribution range; Boundary layer state assessment steps: Utilize a deep learning-based prediction model combined with traditional theoretical models to calculate boundary layer thickness, displacement thickness, momentum thickness, and transition position parameters, and predict the future state of the boundary layer; Control decision-making steps: Adopting a fusion algorithm of hierarchical reinforcement learning and model predictive control, combined with an adaptive exploration-exploitation strategy, to determine the regulation strategy while considering multi-objective optimization; Adjustment execution steps: The adjustment action is driven by the micro-electromechanical system, the control signal is transmitted using coded modulation technology, and the actuators work together; Model validation and optimization steps: real-time comparison of pre- and post-adjustment data with model results, use of Bayesian model averaging to fuse and evaluate different models, and use of online learning algorithms to update the model; Human-computer interaction and remote monitoring steps: Human-computer interaction is achieved through virtual reality and augmented reality technologies, remote monitoring and control are supported, and the system is equipped with early warning and decision-making assistance functions.

9. The method of real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 8, characterized in that: Also includes: Environmental perception and compensation steps: Real-time monitoring of the external meteorological conditions of the wind tunnel, and the use of fuzzy control algorithms to compensate and adjust the boundary layer control strategy according to changes in environmental parameters.

10. The method of real-time adaptive adjustment system for wind tunnel boundary layer control according to claim 8, characterized in that: Also includes: Sensor fault diagnosis and fault tolerance steps: integrate and analyze sensor data and perform residual detection to determine sensor faults, use Kalman filtering to reconstruct faulty sensor data, and adjust the sampling and fusion strategies of each sensor.

Images