Deformed wing state inversion system

Abstract

The invention discloses a deformed wing state inversion system, and aims to solve the problems that the conventional deformed wing is low in implementation of sensing a flight environment and a self-load state, needs a large number of computing resources and the like. The system integrates various sensors, a data processing unit, an inversion module, an autonomous decision-making module and the like, load data of key point positions on a wing are obtained by using the sensors, the load condition of the whole wing surface is inversely predicted by an inversion model with learnability after the load data is preprocessed by the data processing unit, and a visual cloud picture is generated. And a driver or an autonomous decision module judges whether wing body deformation is carried out, so that the flight efficiency is optimized or the wings are prevented from being damaged. The implementation and efficiency of deformation wing state inversion are improved, and an airfoil surface load inversion system with learnability and adaptability is provided and constructed.

Description

Technical Field

[0001] The present invention relates to a method for inverting the load state of a morphing wing in a short time, belonging to the field of aircraft design. Background Art

[0002] In the field of aviation, morphing wing technology has received extensive attention because it can dynamically adjust the aerodynamic characteristics of an aircraft according to flight missions and environmental changes. Traditional fixed-wing aircraft need to be optimized according to specific flight conditions during design, which limits their performance in multilateral environments. In contrast, morphing wing aircraft can adapt to different flight environments by changing the wing geometry, thereby improving flight efficiency, increasing range, improving maneuverability, and reducing fuel consumption.

[0003] The methods for morphing wings to perceive flight environments and structural states are mainly divided into two categories. One is long-distance measurement using external devices such as stereo cameras and scanners, and the other is contact measurement by attaching or embedding sensors on the wing. The operability and portability of long-distance measurement are poor and cannot meet the real-time perception requirements during flight. Contact measurement relies on complex sensor systems and a large amount of computing resources, which not only increases costs but also reduces the feasibility of practical applications.

[0004] In summary, there is an urgent need to develop an efficient, low-cost load state inversion method applicable to morphing wings. Summary of the Invention

[0005] The technical problem to be solved by the present invention is: overcoming the deficiencies of the prior art, providing a morphing wing state inversion system, solving the problems of low practicability in perceiving flight environments and its own load states and high computational resource requirements of existing morphing wings, and realizing the monitoring and inversion of morphing wing states.

[0006] The technical solution of the present invention is:

[0007] A morphing wing state inversion system, comprising: a sensor, a data processing unit, an inversion module, and an autonomous decision-making module;

[0008] The sensor collects load data on the wing and transmits it to the data processing unit. After preprocessing, the data processing unit provides the data to the inversion module. The inversion module feeds back the inverted wing surface load results to the data processing unit and the autonomous decision-making module. The data processing unit converts the feedback data and outputs a visualized result, which is transmitted to the aircraft. The autonomous decision-making module determines whether the wing changes to a shape with higher flight efficiency or prevents itself from being damaged according to the inverted wing surface load results, and provides the judgment results to the aircraft.

[0009] Further, the sensors include pressure sensors, strain sensors, and temperature sensors, which are arranged at key structural or mechanism positions of the deformable wing, including the wing root, leading edge, wing tip, and deformation mechanism.

[0010] Further, the pressure sensor is used to measure the pressure change of the airflow on or around the surface of the deformable wing to provide aerodynamic load data; the strain sensor is used to measure the strain caused by the aerodynamic load; and the temperature sensor is used to measure the temperature of the surface of the deformable wing to provide aerodynamic heat load.

[0011] Further, the data processing unit performs preprocessing, specifically referring to cleaning, denoising, and normalizing the collected raw data.

[0012] Further, the inversion module performs wing surface inversion on the preprocessed data to obtain the wing surface load result. Specifically, the state inversion method based on the BP neural network is used for wing surface inversion.

[0013] Further, the data processing unit converts the feedback data and outputs it as a visualized load distribution cloud map.

[0014] Further, the autonomous decision-making module determines whether the wing changes to a shape with higher flight efficiency or prevents itself from being damaged according to the inverted wing surface load result. The specific operation is as follows: when the aerodynamic load / strain is less than the preset lower threshold, the windward surface of the wing is increased; when the aerodynamic load / strain is greater than the preset upper threshold, the structure mechanism may be damaged, and the wing is appropriately retracted or folded.

[0015] The beneficial effects of the present invention compared with the prior art are as follows:

[0016] (1) The present invention simplifies the sensor layout, reduces the number of sensors, and lowers the system cost and complexity. The key positions, such as the wing root, leading edge, wing tip, and deformation mechanism, can be determined by analyzing the structure and loading characteristics of the deformable wing, so as to determine the final sensor arrangement position, and realize the acquisition of data at key positions on the wing surface and the accurate inversion of the wing surface load state.

[0017] (2) The present invention establishes an inversion model and uses intelligent algorithms to learn the load transfer mode of the deformable wing, improving the accuracy and efficiency of state inversion.

[0018] (3) The inversion model in the present invention can learn deformable wings with different shapes and deformation methods, has a certain adaptability and universality, and can be continuously iterated according to data to improve the inversion accuracy and speed.

[0019] (4) The inversion module in the present invention is designed with reserved interfaces for accessing autonomous intelligent decision-making, thereby realizing fully autonomous wing surface inversion and deformation and improving flight efficiency. Description of the Drawings

[0020] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0021] Figure 2 This is a schematic diagram of the simplified sensor arrangement of the variable wing;

[0022] Figure 3 This is a monitoring inversion flowchart;

[0023] Figure 4 This is an inversion training flowchart. Specific Embodiments

[0024] The following further describes in detail the specific embodiments of the present invention with reference to the accompanying drawings.

[0025] The present invention proposes a variable wing state inversion system, aiming to solve the problems of low feasibility of existing variable wings in perceiving flight environment and their own load states, and high computational resource requirements. The system integrates various sensors, a data processing unit, an inversion module, an autonomous decision-making module, etc. By using sensors to obtain load data at key points on the wing, after preprocessing by the data processing unit, the load conditions of the entire wing surface are inversely predicted by a learning-based inversion model, and a visualization cloud map is generated, which is handed over to the pilot or the autonomous decision-making module to determine whether to perform wing-body deformation to optimize flight efficiency. The present invention improves the feasibility and efficiency of variable wing state inversion and provides a wing surface load inversion system with learnability and adaptability.

[0026] As Figure 1 shown, the present invention proposes a variable wing state inversion system, including: sensors, a data processing unit, an inversion module, and an autonomous decision-making module;

[0027] The sensors collect load data on the wing and transmit it to the data processing unit. After preprocessing by the data processing unit, the data is provided to the inversion module. The inversion module feeds back the inversed wing surface load results to the data processing unit and the autonomous decision-making module. The data processing unit converts the feedback data and outputs it as a visualization result, which is transmitted to the aircraft; the autonomous decision-making module determines whether the wing changes to a shape with higher flight efficiency or prevents itself from being damaged based on the inversed wing surface load results, and provides the judgment result to the aircraft.

[0028] As Figure 2The figure shows a simplified schematic diagram of the sensor arrangement on a morphing wing. The sensor arrangement strategy is studied to achieve optimal monitoring results with a minimized number of sensors. The sensors are arranged at key structural or mechanism positions of the morphing wing, such as the wing root, leading edge, wing tip, and morphing mechanism. The sensors can be divided into pressure sensors, strain sensors, and temperature sensors. Among them, pressure sensors can measure the pressure changes on the surface of the morphing wing or in the surrounding airflow to provide aerodynamic load data; strain sensors can measure the strain caused by aerodynamic loads or other external forces; temperature sensors can measure the temperature on the surface of the morphing wing to provide aerodynamic heat loads. With a limited number of sensors, it is sufficient to collect key data representing the load state of the entire wing surface, while reducing the weight and cost of the aircraft.

[0029] Inversion module: An inversion module that describes the load transfer mode of the morphing wing is established. Through machine learning, the model is trained with a large amount of flight data so that it can learn the load transfer mode on the wing, enabling the inversion of the wing surface load state based on the data collected by the sensors with a simplified layout. In addition, intelligent algorithms can be trained according to data in specific scenarios to improve their performance in extreme or sudden flight situations.

[0030] The present invention uses a state inversion method based on a BP neural network for inversion. The number of input layer and output layer units of the BP neural network is related to the specific data input and output, while the number of hidden layers and the number of neurons in each hidden layer directly affect the training rate and accuracy of the network. Increasing the number of hidden layers and the number of neurons in each hidden layer will, on the one hand, improve the results of the training model, but on the other hand, it will also cause a rapid increase in the training duration and may also result in overfitting. Therefore, a suitable stopping criterion is constructed to train a suitable network result, and a suitable number of hidden layers and the number of neurons in each layer are selected.

[0031] The construction and training of the model are mainly completed through importing data, extracting input features and target variables, dividing the training set and test set, normalizing processing, constructing a BP neural network, configuring network parameters, training the BP neural network, using test samples to predict through the trained model, and denormalizing the prediction results and calculating the error to complete the construction of the inversion model and the analysis of the model error accuracy.

[0032] Data processing unit: It converts the feedback data provided by the inversion module, outputs a visualized load distribution contour map, and transmits it to the aircraft. For example, its x-axis and y-axis are spanwise and chordwise coordinates, and different colors are used to represent different temperature levels. Design a high-performance data processing unit to process sensor data, evaluate the load inversion results, and then generate reference feedback information. Preprocess the sensor data, including cleaning, denoising, and standardization, to improve data quality. At the same time, the data processing unit has an interface with the intelligent algorithm, responsible for inputting the preprocessed data into the algorithm (state inversion method based on BP neural network), and converting the results inverted by the algorithm into available load state information.

[0033] Such as Figure 3 The monitoring inversion flowchart is shown as follows, which demonstrates the basic process of the deformed wing load state inversion.

[0034] (1) Start: The inversion system starts to work;

[0035] (2) Data acquisition: Collect data from various sensors during flight;

[0036] (3) Determine whether data acquisition is complete: Check whether the collected data meets the requirements of the inversion algorithm for processing;

[0037] (4) Data preprocessing: Clean, denoise, and standardize the collected raw data;

[0038] (5) Wing surface inversion: Use the trained wing surface inversion model to analyze and simulate the processed data.

[0039] (6) Output of inversion results: Output the inversion results to form a complete wing surface load contour map.

[0040] (7) End: End the inversion process. You can start the next inversion according to the requirements, or set the inversion frequency for automatic inversion.

[0041] Such as Figure 4 The inversion training flowchart is shown as follows, which demonstrates the basic process of the algorithm training of the inversion module.

[0042] (1) First, collect a large number of state parameters at various locations on the inner and outer surfaces of the wing under different force and thermal load environments

[0043] (2) Then use this data to train the neural network model. By adjusting the number of input layers, hidden layers, and output layers, and training the weights between layers, a state mapping relationship between the preset input measurement feature point states and the states at other positions is constructed;

[0044] (3) Obtain the inversion accuracy by comparing the inversion results with the reference data;

[0045] (4) The neural network model can be saved after the fitting precision error reaches the set requirements, that is, the training is completed.

[0046] The parts not detailed in the present invention are common general knowledge to those skilled in the art.

Claims

1. A morphing wing state inversion system, characterized in that Including: A sensor, a data processing unit, an inversion module, and an autonomous decision-making module; The sensor collects the load data on the wing and transmits it to the data processing unit. After preprocessing, the data processing unit provides the data to the inversion module. The inversion module feeds back the wing surface load results obtained by inversion to the data processing unit and the autonomous decision-making module. The data processing unit converts the feedback data and outputs it into a visual result, which is transmitted to the aircraft; Based on the wing surface load results obtained by inversion, the autonomous decision-making module determines whether the wing changes to a shape with higher flight efficiency or prevents itself from being damaged, and provides the judgment result to the aircraft.

2. The morphing wing state inversion system according to claim 1, characterized in that: The sensor includes a pressure sensor, a strain sensor, and a temperature sensor, which are arranged at the key structures or mechanism positions of the deformable wing, including the wing root, the leading edge, the wing tip, and the deformation mechanism.

3. The morphing wing state inversion system according to claim 2, characterized in that: The pressure sensor is used to measure the pressure change of the airflow on or around the surface of the deformable wing and provide aerodynamic load data; the strain sensor is used to measure the strain caused by the aerodynamic load; the temperature sensor is used to measure the temperature of the surface of the deformable wing and provide aerodynamic heat load.

4. A morphing wing state inversion system according to claim 1, characterized in that: The data processing unit performs preprocessing, specifically referring to cleaning, denoising, and normalizing the collected raw data.

5. A deformation wing state inversion system according to claim 1, characterized in that: The inversion module performs wing surface inversion on the preprocessed data to obtain the wing surface load results. Specifically: the state inversion method based on the BP neural network is used for wing surface inversion.

6. The morphing wing state inversion system according to claim 5, characterized in that: The data processing unit converts the feedback data and outputs it into a visual load distribution cloud map.

7. A morphing wing state inversion system according to claim 5, characterized in that: Based on the wing surface load results obtained by inversion, the autonomous decision-making module determines whether the wing changes to a shape with higher flight efficiency or prevents itself from being damaged. The specific operation is as follows: if the aerodynamic load / strain is less than the preset lower threshold, the windward surface of the wing is increased; if the aerodynamic load / strain is greater than the preset upper threshold, the structural mechanism may be damaged, and the wing is appropriately retracted or folded.

Images