A wing monitoring system for an aircraft and an aircraft

Abstract

The application relates to the field of aviation technology, in particular to a wing monitoring system of an aircraft and the aircraft, which is used for high-precision and real-time sensing of vibration and deformation of a wing and accurate evaluation of the state of the wing. The wing monitoring system comprises a sensor system and a monitoring platform in communication connection, wherein the sensor system comprises a TENG module, an EMG module and a PENG module, is installed on the wing of the aircraft and is used for outputting a first electric signal and a second electric signal representing vibration information of the wing and a third electric signal representing deformation information of the wing; the monitoring platform is used for acquiring the first electric signal, the second electric signal and the third electric signal, and analyzing the first electric signal, the second electric signal and the third electric signal based on a pre-trained neural network model to predict the vibration information and the deformation amount of the wing.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of aviation technology, in particular to a wing monitoring system of an aircraft and the aircraft. BACKGROUND

[0002] With the development of modern aircraft towards higher speed, lighter structure and higher task load, the load-bearing components such as wings are subjected to continuous complex aerodynamic loads during flight, which are prone to structural vibration, deformation and fatigue damage problems. The aeroelastic response of the wing not only directly affects the lift and stability, but also may induce stall, material micro-crack propagation and even structural disintegration and other serious consequences.

[0003] Therefore, establishing a high-precision and real-time wing vibration and deformation monitoring mechanism is a key prerequisite for realizing structural health assessment and safety warning. SUMMARY

[0004] The present application discloses a wing monitoring system of an aircraft and the aircraft, which can accurately assess the state of the wing by high-precision and real-time sensing of the vibration and deformation of the wing.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] In a first aspect, the present application provides a wing monitoring system of an aircraft, comprising a sensor system and a monitoring platform in communication connection, wherein,

[0007] The sensor system comprises a triboelectric nanogenerator (TENG) module, an electromagnetic generator (EMG) module and a piezoelectric nanogenerator (PENG) module, which are installed on the wing of the aircraft and used to output a first electric signal and a second electric signal representing the vibration information of the wing, and a third electric signal representing the deformation information of the wing, wherein the TENG module and the EMG module are respectively used to output the first electric signal and the second electric signal, and the PENG module is used to output the third electric signal;

[0008] The monitoring platform is used to acquire the first electric signal, the second electric signal and the third electric signal, and based on a pre-trained neural network model, to analyze the first electric signal, the second electric signal and the third electric signal, and to predict the vibration information and the deformation of the wing.

[0009] The wing monitoring system of the aircraft provided by the embodiments of the present application can accurately distinguish the vibration information and deformation information of the wing in a complex dynamic environment without an external power supply, realize high-precision and real-time sensing of the vibration and deformation of the wing, and simultaneously, based on a pre-trained neural network model, a monitoring platform analyzes the electrical signals output by the sensor system, predicts the vibration information and deformation of the wing, and realizes remote digital evaluation of the dynamic response of the wing structure.

[0010] Further, the pre-trained neural network model is generated by training in the following manner:

[0011] Obtain training sample data, each of which includes a first sample electrical signal, a second sample electrical signal, a third sample electrical signal, sample vibration information, and a sample deformation;

[0012] For any training sample data, the following operations are performed until the training sample data are all processed or the prediction accuracy reaches a preset requirement, and the neural network model is generated:

[0013] Pretreat the first sample electrical signal, the second sample electrical signal, and the third sample electrical signal to obtain three groups of time series signals of a preset length;

[0014] Process the three groups of time series signals of a preset length using a preset convolution kernel, extract the time characteristics of each group of time series signals according to the target weight of each sample electrical signal, and the time characteristics represent the relationship between the signal parameters and time in the time series signal;

[0015] Splice the time characteristics of the groups of time series signals to obtain spliced features, extract dynamic evolution features from the spliced features through a bidirectional long short-term memory (LSTM) layer, and the dynamic evolution features represent the change relationship of the signal with time in the time series signal;

[0016] Output estimated vibration information and estimated deformation based on the dynamic evolution features, calculate the prediction error based on the estimated vibration information, estimated deformation, sample vibration information, and sample deformation through a preset loss function, and adjust the target weight of each sample electrical signal through a back propagation algorithm, wherein the preset loss function includes a regularization loss term based on a vibration physical formula.

[0017] Further, the regularization loss term based on the vibration physical formula is as follows:

[0018] Formula One

[0019] wherein, MSE is mean square error, represents the vibration acceleration in the estimated vibration information, represents the vibration frequency in the estimated vibration information, represents the vibration amplitude in the estimated vibration information.

[0020] Further, the preset convolution kernel is three groups of 1D convolution kernels with different scales.

[0021] Further, the monitoring platform is further configured to display the change curves of the first electrical signal, the second electrical signal, and the third electrical signal.

[0022] Further, the monitoring platform is further configured to trigger an alarm prompt when it is determined that the vibration information and / or the deformation amount meet a preset alarm condition.

[0023] Further, the TENG module comprises a first electrode, a second electrode, a permanent magnet, a plurality of support columns, and a first load component, wherein,

[0024] The plurality of support columns are arranged between the first electrode and the second electrode, and a cavity is formed between the plurality of support columns. The permanent magnet is arranged in the cavity. A first material is arranged on the surface of the permanent magnet. A second material is arranged on the side of the first electrode opposite to the permanent magnet. A second material or a third material is arranged on the side of the second electrode opposite to the permanent magnet. The electronegativity of the first material is different from that of the second material, and the electronegativity of the first material is different from that of the third material. The first end of the first load component is connected to the first electrode, and the second end is connected to the second electrode.

[0025] Further, the EMG module comprises a coil, a non-magnetic shell, and a second load component. The non-magnetic shell surrounds the permanent magnet and is arranged in the cavity. The coil is wound on the non-magnetic shell. The first end of the second load component is connected to one end of the coil, and the second end is connected to the other end of the coil.

[0026] Further, the PENG module comprises a first piezoelectric unit, a second piezoelectric unit, and a differential circuit. The first piezoelectric unit is fixed to the surface of the wing through silica gel. The second piezoelectric unit is rigidly fixed to the first piezoelectric unit. The input ends of the differential circuit are connected to the first piezoelectric unit and the second piezoelectric unit, respectively.

[0027] In a second aspect, the embodiments of the present application provide an aircraft, comprising a fuselage, a wing, and the wing monitoring system provided by the first aspect of the embodiments of the present application.

[0028] The wing monitoring system is used for sensing the vibration information and deformation information of the wing and predicting the vibration information and deformation amount of the wing. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 An architecture schematic diagram of a wing monitoring system of an aircraft is provided for the embodiments of the present application.

[0030] Figure 2 A structure schematic diagram of a TENG module is provided for the embodiments of the present application.

[0031] Figure 3 A structure schematic diagram of another TENG module is provided for the embodiments of the present application.

[0032] Figure 4 A structure schematic diagram of an EMG module is provided for the embodiments of the present application.

[0033] Figure 5 A working principle schematic diagram of the TENG module and the EMG module is provided for the embodiments of the present application.

[0034] Figure 6 A structure schematic diagram of a PENG module is provided for the embodiments of the present application.

[0035] Figure 7 A working principle schematic diagram of the PENG module is provided for the embodiments of the present application.

[0036] Figure 8 An integrated structure schematic diagram of a sensor system is provided for the embodiments of the present application.

[0037] Figure 9 Another integrated structure schematic diagram of a sensor system is provided for the embodiments of the present application.

[0038] Figure 10 A signal response characteristic schematic diagram of the TENG module in a process of increasing wind speed is provided for the embodiments of the present application.

[0039] Figure 11 A signal response characteristic schematic diagram of the EMG module in a process of increasing wind speed is provided for the embodiments of the present application.

[0040] Figure 12 A signal response characteristic schematic diagram of the PENG module in a process of increasing deformation amount is provided for the embodiments of the present application.

[0041] Figure 13 A principle schematic diagram of a training process of a neural network model is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0042] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0043] The application scenarios described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those of ordinary skill in the art can know that, with the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems. In the description of the present application, unless otherwise specified, the meaning of “multiple” is two or more.

[0044] Before introducing the wing monitoring system of the aircraft and the aircraft provided by the embodiments of the present application, in order to facilitate understanding, the technical background of the embodiments of the present application is first introduced in detail.

[0045] With the development of modern aircraft towards higher speed, lighter structure and higher task load, the load-bearing components such as wings are subjected to continuous complex aerodynamic loads during flight, which is prone to cause structural vibration, deformation and fatigue damage problems. The aeroelastic response of the wing not only directly affects the lift and stability, but also can induce stall, material micro-crack propagation and even structural disintegration and other serious consequences.

[0046] Therefore, establishing a high-precision and real-time wing vibration and deformation monitoring mechanism is a key prerequisite for realizing structural health assessment and safety warning.

[0047] In view of this, the wing monitoring system of the aircraft and the aircraft provided in the embodiments of the present application. The wing monitoring system of the aircraft provided by the embodiments of the present application can accurately distinguish the vibration information and deformation information of the wing in a complex dynamic environment without external power supply, by using multiple self-powered modules to perceive the vibration and deformation of the wing, thereby realizing high-precision and real-time perception of the vibration and deformation of the wing. At the same time, based on the pre-trained neural network model, the monitoring platform analyzes the electrical signals output by the sensor system, predicts the vibration information and deformation of the wing, and realizes remote digital evaluation of the dynamic response of the wing structure.

[0048] It should be noted that the wing monitoring system of the aircraft provided by the embodiments of the present application can be applied to various types of aircraft, such as fixed-wing aircraft, rotary-wing aircraft, unmanned aerial vehicles, etc., and can also be extended to real-time state perception and health assessment of flexible structures such as wind turbine blades, bridges, track beams and ship masts.

[0049] After introducing the background of the embodiments of the present application, the overall concept of the embodiments of the present application is described below.

[0050] Currently, the aircraft wing is subjected to the vibration and deformation coupling behavior under the action of complex aerodynamic load in long-term service, which is not only the fundamental reason for material fatigue, structural crack and aerodynamic performance degradation, but also directly threatens flight safety and aircraft life evaluation. However, the existing scheme still has the following technical problems to be broken through in terms of sensing principle, structure design, energy supply and data processing capability:

[0051] 1. The signal modal aliasing is serious, and it is difficult to realize effective decoupling identification of vibration and deformation.

[0052] The existing sensor can usually only collect a single physical modal signal, and is easily disturbed by the overlapping structure response, and it is difficult to accurately distinguish the vibration and deformation effects in a complex dynamic environment.

[0053] 2. Strong structural invasiveness and significant disturbance, difficult to meet the requirement of high aerodynamic performance.

[0054] The existing sensor needs to be embedded or externally attached and installed destructively, which destroys the wing surface integrity and causes boundary layer separation, vortex excitation and aerodynamic efficiency decline.

[0055] 3. Relies on external power supply system, difficult to realize distributed deployment and long-term monitoring.

[0056] The existing sensor needs external power supply or centralized wiring, which increases the system complexity, weight and reliability risk.

[0057] 4. The signal is seriously disturbed, and it is difficult to realize high stability and reliability monitoring.

[0058] In the high flow rate and strong vibration environment, the existing contact type sensor has problems such as signal drift, adhesion failure, response lag, etc.

[0059] 5. Strong dependence on artificial, insufficient intelligent recognition accuracy and generalization ability.

[0060] Most traditional schemes rely on manual setting of parameters (such as Fast Fourier Transform (FFT) window length, filter frequency band, etc.), which are difficult to adapt to nonlinear and non-stationary vibration-deformation coupling response.

[0061] To solve the above technical problems, the wing monitoring system of the aircraft provided in the embodiments of the present application integrates a TENG module, an EMG module and a PENG module in the sensor system, constructs a magnetic suspension vibration detection structure and a soft and hard integrated differential deformation detection structure, and optimizes the aerodynamic performance through a streamlined shell and a vortex generator, and deploys a neural network model fusing physical constraint learning and attention mechanism in the monitoring platform to improve the signal recognition capability and accurately evaluate the health status of the wing.

[0062] In the wing monitoring system of the aircraft provided in the embodiments of the present application, as for the sensor system, vibration information is perceived through the TENG module and the EMG module, and deformation information is perceived through the double-layer piezoelectric unit in the PENG module + differential decoupling mechanism, realizing physical differentiation and high-precision identification of coupled signals under in-situ perception conditions; the TENG module and the EMG module adopt a magnetic suspension triboelectric-electromagnetic dual-power generation structure, combined with a flexible piezoelectric film, to realize self-powered operation, support multi-point distribution and light deployment, and significantly improve system integration and maintenance convenience; moreover, the magnetic suspension structure avoids mechanical wear and enhances fatigue resistance, and its stability is verified through 24-hour wind tunnel fatigue test, and the TENG and PENG signal attenuation rates are better than the existing technical level; finally, the TENG module, the EMG module and the PENG module are integrated in the same shell, and through streamlined shell and vortex generator integrated design, they can be installed at a suitable position of the wing to realize optimal aerodynamic performance layout, minimize resistance and improve lift, and take into account monitoring function and aerodynamic stability from the source.

[0063] As for the monitoring platform, the neural network model fusing physical constraint learning and attention mechanism is deployed in the monitoring platform to process the electrical signals output by the sensor system, realizing end-to-end, high-robustness and transferable signal recognition capability on the basis of ensuring prediction accuracy.

[0064] After introducing the background technology and overall concept of the embodiments of the present application, the wing monitoring system of the aircraft and the aircraft provided in the embodiments of the present application will be described in detail in combination with specific embodiments.

[0065] Referring to Figure 1 Fig. 1 is an architecture schematic diagram of a wing monitoring system of an aircraft in the embodiments of the present application, which comprises a sensor system 10 and a monitoring platform 11 in communication connection.

[0066] The sensor system 10 includes a TENG module, an EMG module, and a PENG module, which are installed on the wing of an aircraft and are used to output a first electrical signal and a second electrical signal representing wing vibration information, and a third electrical signal representing wing deformation information. The TENG module and the EMG module are used to output the first electrical signal and the second electrical signal, respectively, and the PENG module is used to output the third electrical signal.

[0067] The monitoring platform 11 is used to obtain the first electrical signal, the second electrical signal, and the third electrical signal, and analyze the first electrical signal, the second electrical signal, and the third electrical signal based on a pre-trained neural network model to predict the vibration information and deformation of the wing.

[0068] Among them, the monitoring platform can be a web-based visualization platform built based on ECharts and jQuery.

[0069] The following describes in detail the structure of each module in the sensor system provided by the embodiment of the present application in conjunction with specific embodiments.

[0070] like Figure 2 As shown, the TENG module includes: a first electrode 21, a second electrode 22, a permanent magnet 23, a plurality of support pillars (not shown in the figure), and a first load component 24. The plurality of support pillars are arranged between the first electrode 21 and the second electrode 22, and a cavity is formed between the plurality of support pillars. The permanent magnet 23 is placed in the cavity. The surface of the permanent magnet 23 is provided with a first material. The side of the first electrode 21 opposite to the permanent magnet 23 is provided with a second material. The side of the second electrode 22 opposite to the permanent magnet 23 is provided with a second material or a third material. The first material and the second material have different electronegativity, and the first material and the third material have different electronegativity. The first end of the first load component 24 is connected to the first electrode 21, and the second end is connected to the second electrode 22.

[0071] Among them, the first material and the second material can be any pair of material combinations with different electronegativity, such as nylon membrane, polyamide PA, polytetrafluoroethylene PTFE, fluorinated ethylene propylene copolymer FEP, polyimide Kapton, polyvinylidene fluoride PVDF, etc. Similarly, the first material and the third material can also be any pair of material combinations with different electronegativity, such as nylon membrane, PA, PTFE, FEP, Kapton, PVDF, etc.

[0072] like Figure 3As shown, the TENG module in the embodiment of the present application adopts a four-pole support column (four support columns 30) equidistant arrangement to support the permanent magnet, and this structure can be extended to other magnetic arrangement schemes, including but not limited to three-pole, six-pole, and even annular permanent magnet array structure. Meanwhile, the permanent magnet in the embodiment of the present application adopts a cylindrical permanent magnet, and of course, it can also be replaced by a spherical, ellipsoidal, or multi-segment laminated structure according to actual vibration response requirements.

[0073] For example, the nylon film is arranged on the surface of the permanent magnet, and the PTFE film is arranged on the side of the first electrode opposite to the permanent magnet and the side of the second electrode opposite to the permanent magnet. The TENG module in the embodiment of the present application is based on magnetic suspension design, adopts an upper and lower symmetrical triboelectric structure, and the nylon film alternately contacts and separates from the upper and lower PTFE films during the up and down movement of the permanent magnet, thereby generating a periodic triboelectric signal. The triboelectric signal is extracted by the first load component, and the triboelectric signal can be used as an electrical signal representing the vibration information of the object.

[0074] In specific implementation, as shown in Figure 4 The EMG module includes a coil 41, a non-magnetic shell, and a second load component. The non-magnetic shell is arranged around the permanent magnet, the non-magnetic shell is arranged in the cavity, the coil 41 is wound on the non-magnetic shell, and the first end of the second load component is connected with one end of the coil, and the second end is connected with the other end of the coil.

[0075] In actual application, the coil in the EMG module can adopt a copper coil, the number of turns of the coil is in the range of 1500±100, which can be adjusted according to the sensitivity requirement, and the coil material can be copper, silver-plated copper wire, alloy wire, or printed conductive ink. In addition, the magnetic flux cutting path can be enhanced by optimizing the structure shell slotting mode, adjusting the support column angle, and designing the spiral magnetic flux guide groove, so as to improve the signal response ability under low amplitude working condition.

[0076] In the EMG module provided in the embodiment of the present application, when the object vibrates, the permanent magnet cuts the magnetic field lines in the vibration, generates an EMG electrical signal according to the Faraday's law of electromagnetic induction, and extracts the EMG electrical signal through the second resistance component, so as to use the EMG electrical signal as an electrical signal representing the vibration information of the object.

[0077] The working principles of the TENG module and the EMG module in the sensor system provided in the embodiment of the present application will be described below. Figure 5 As shown in

[0078] As shown in Figure 5 The initial state of the TENG module and the EMG module is as shown in Figure 5 When the object vibrates, the permanent magnet generates periodic vertical movement under the vibration excitation, and the permanent magnet cuts the magnetic field lines in the vibration, thereby generating an EMG electrical signal according to the Faraday's law of electromagnetic induction. Figure 5 Figure 5 ​In the middle (B) state, the magnet cuts the copper coil to generate electromagnetic induction current, the direction of which is determined by the change of the magnetic field. Figure 5 Medium (B) state to Figure 5 In the middle (D) state, the nylon membrane on the magnet and the PTFE on the cylinder wall contact and separate, generating frictional charge transfer and the formation of triboelectric current. The triboelectric current and electromagnetically induced current together characterize the vibration, and their generation process is directly related to the mechanical stimulus.

[0079] In specific implementation, the PENG module provided in the embodiment of the present application is as follows: Figure 6 As shown, it includes: a first piezoelectric unit 61, a second piezoelectric unit 62 and a differential circuit 63. The first piezoelectric unit 61 is fixed to the wing surface through silicone, the second piezoelectric unit 62 is rigidly fixed on the first piezoelectric unit 61, and the input end of the differential circuit 63 is connected to the first piezoelectric unit 61 and the second piezoelectric unit 62 respectively.

[0080] The first piezoelectric unit 61 and the second piezoelectric unit 62 can each be formed of two electrode layers and a PVDF film layer, with the PVDF film layer sandwiched between the two electrode layers. The electrode layer materials can be copper, aluminum, gold, silver, carbon nano-ink, etc. The PVDF film layer can also be replaced with other piezoelectric materials such as PZT piezoelectric ceramics, PVDF-TrFE composite films, BaTiO microfilms, ZnO nano-arrays, etc.

[0081] In practical applications, the first piezoelectric unit 61 can be embedded in the silicone layer to sense the comprehensive strain of the object, and the second piezoelectric unit 62 can be installed on the rigid upper shell 64 to respond only to the vibration of the object. The electrical signals generated by the two are decoupled through the differential circuit 63 to extract the deformation signal.

[0082] It should be noted that the PENG module in the embodiment of the present application includes two layers of piezoelectric units. In other embodiments of the present application, the PENG module can also be expanded to a structure with three or more layers to achieve multi-layer sensitivity enhancement. The embodiment of the present application does not limit the number of piezoelectric units in the PENG module.

[0083] The following combination Figure 7 The working principle of the PENG module in the sensor system provided in the embodiment of the present application is described.

[0084] like Figure 7 As shown, the initial state of the PENG module is as follows Figure 7 As shown in (A), a double-layer piezoelectric structure is used to synchronously collect deformation and vibration signals. The upper piezoelectric unit is installed on a rigid base (for example, a PLA rigid base) and only responds to vibration. The lower piezoelectric unit is embedded in flexible silicone and is sensitive to both vibration and deformation, generating a composite current. Pure deformation signals can be extracted through differential calculations. As vibration and deformation occur, the distributed current in the PENG module is as follows:Figure 7 As shown in (B), (C), and (D), the final output is current i PENG1 and i PENG2 differential signal.

[0085] In practical applications, the TENG module and the EMG module are taken as a whole, and they and the PENG module can be deployed in a distributed manner or integrated in a shell.

[0086] When the TENG module, EMG module, and PENG module are integrated in the same shell, in order to improve the aerodynamic compatibility of the sensor system in a high-speed flight environment, the embodiment of the present application implements aerodynamic optimization of the sensor system shell. Specifically, the shell can adopt a streamlined structure, and a vortex generator structure and / or an airflow guide channel can also be set on the shell. The airflow attachment and boundary layer control are optimized through CFD simulation to reduce the airflow separation area at high angles of attack.

[0087] In practice, a streamlined shell is used for aerodynamic optimization. Multiple front and rear guide holes, vortex-generating ridges, and curved contraction channels can be designed to match flow fields at varying angles of attack. This geometric arrangement can be expanded to include asymmetric hole patterns, variable-angle adjustable structures, or foldable shell designs. Shell materials include, but are not limited to, polycarbonate, PETG, carbon fiber reinforced composites, and aviation-grade PEEK.

[0088] like Figure 8 As shown, the TENG module, EMG module, and PENG module are integrated in the same shell 80. The TENG module and EMG module adopt a magnetic suspension structure 81 with multiple support columns. Specifically, four equally spaced magnetic columns can be used to support the cylindrical permanent magnet to achieve contactless suspension vibration response; the PENG module serves as the base plate 82. Specifically, a soft-hard composite base plate can be used with an embedded double-layer PVDF piezoelectric film, one of which is fixed by silicone to simultaneously sense vibration and deformation, and the other is rigidly fixed to the base plate and only responds to vibration, thereby achieving signal decoupling through a differential structure.

[0089] The following describes the internal structure of the TENG module, EMG module, and PENG module when they are integrated into the same housing, using a top view. Figure 9 As shown, a cavity is formed between four equally spaced magnetic columns 90 , in which a cylindrical permanent magnet can be placed. A non-magnetic shell 91 is also provided in the cavity, on which a coil can be wound.

[0090] The following combination Figure 10- Figure 12 , the wind tunnel test verification results of each module in the sensor system provided in the embodiment of the present application are described.

[0091] like Figure 10 As shown, Figure 10The signal response characteristics of the TENG module during the process of increasing wind speed are shown. From 9.6 m / s to 36.9 m / s, the signal strength is significantly improved, indicating that the sensor system is sensitive to changes in vibration state.

[0092] As shown in Figure 11 , Figure 11 The signal response characteristics of the EMG module during the process of increasing wind speed are shown. From 9.6 m / s to 36.9 m / s, the signal strength is significantly improved, indicating that the sensor system is sensitive to changes in vibration state.

[0093] As shown in Figure 12 , Figure 12 The voltage response of the PENG module during the process of changing the deformation from 0 to 10 mm is shown. The signal amplitude shows a significant linear growth, verifying the deformation sensing ability of the PVDF film and the reliability of the decoupling structure.

[0094] The above describes the sensor system in the wing monitoring system of the aircraft according to the embodiments of the present application, and the monitoring platform is described in detail below.

[0095] In the monitoring platform, the multi-scale physical perception attention network is constructed to perform regression inference training on vibration information and deformation, and a frequency-deformation joint training mechanism is adopted, so that the inference accuracy of the neural network model is improved, and the inference curve is highly consistent with the real signal.

[0096] In training the neural network model, first, training sample data is obtained, each training sample data at least including a first sample electric signal, a second sample electric signal, a third sample electric signal, sample vibration information, and sample deformation, and the specific training process of the neural network model is as shown in Figure 13 , the following operations are performed for any training sample data until the training sample data are all processed or the prediction accuracy reaches the preset requirement, and the neural network model is generated:

[0097] Operation 1: Preprocessing. That is, the first sample electric signal (TENG signal), the second sample electric signal (EMG signal), and the third sample electric signal (PENG signal) are preprocessed to obtain three groups of time series signals with a preset length.

[0098] The preprocessing includes normalization processing, bandpass filtering, etc., to ensure the consistency and stability of the model input data. The specific normalization processing and bandpass filtering processing can adopt the manner in the related art, and the preset length can be set according to experience, for example, the preset length is 15 seconds.

[0099] Operation 2: feature extraction. Specifically, first, the three groups of preset length time series signals are processed respectively by using preset convolution kernels, and the time features of each group of time series signals are extracted according to the target weight of each sample electric signal, so as to capture the short-term and long-term physical response mode, and fully express the relationship between the parameters (such as frequency) of the signal in each electric signal and time. Among them, the time feature represents the relationship between the signal parameter and time in the time series signal.

[0100] Among them, the preset convolution kernel is a 1-dimensional convolution kernel with three different scales, for example, the size of the preset convolution kernel is 10, 50 and 100, the first sample electric signal is processed by using the convolution kernel with the size of 10, the second sample electric signal is processed by using the convolution kernel with the size of 50, and the third sample electric signal is processed by using the convolution kernel with the size of 100.

[0101] Secondly, the time features of each group of time series signals are spliced to obtain spliced features, and dynamic evolution features are extracted from the spliced features through a bidirectional LSTM layer to improve the perception ability and expression dimension of the model to the time series information. Among them, the dynamic evolution feature represents the change relationship of the signal with time in the time series signal.

[0102] In specific implementation, the bidirectional LSTM network layer can be realized in a manner in the related art, which is not limited in the present application.

[0103] Operation 3: prediction. Specifically, the estimated vibration information and the estimated deformation variable are output based on the dynamic evolution features, the prediction error is calculated by using a preset loss function based on the estimated vibration information, the estimated deformation variable, the sample vibration information and the sample deformation variable, and the target weight of each sample electric signal is adjusted by using a back propagation algorithm, wherein the preset loss function includes a regularization loss term based on a vibration physical formula.

[0104] The regularization loss term based on the vibration physical formula is as follows:

[0105] Formula 1

[0106] Among them, is the regularization loss term, MSE is the mean square error, represents the vibration acceleration in the estimated vibration information, represents the vibration frequency in the estimated vibration information, represents the vibration amplitude in the estimated vibration information.

[0107] It should be noted that the above is the neural network model of the embodiments of the present application. In other embodiments of the present application, the deep learning inference model can be extended to other structural forms of models, including but not limited to the Transformer series, the time convolution network, the lightweight hybrid network (such as MobileNet+LSTM), etc. The input signal can also be extended to image signals, vibration sound signals, or structural stress spectrum data, etc. according to the application scenario; and the predicted value can also be extended from the vibration information and the deformation variable to the fatigue accumulation value, the crack risk score, and the structural health index, etc.

[0108] In practical applications, after the neural network model is generated, the neural network model can be deployed in the monitoring platform to analyze the first electric signal, the second electric signal, and the third electric signal after the monitoring platform obtains the first electric signal, the second electric signal, and the third electric signal output by the sensor system, so as to predict the vibration information and the deformation variable of the wing.

[0109] It should be noted that when the sensor system transmits the first electric signal, the second electric signal, and the third electric signal to the monitoring platform by wireless transmission, the electric signal needs to be converted into a digital signal, and a specific frequency band is used for transmission, such as using a 433 MHz frequency band for transmission. This frequency band has the advantages of strong penetration, high anti-interference performance, and long transmission distance (up to 300 m), and is suitable for deployment in complex flight platform environments.

[0110] In some embodiments, the monitoring platform provided by the embodiments of the present application can also implement the following functions, such as displaying the change curve of the first electric signal, the second electric signal, and the third electric signal; providing FFT spectrum and power distribution graphs; providing historical data query and export services; dynamically displaying the health index of the wing structure; and triggering an alarm prompt when the vibration information and / or the deformation variable meet the preset alarm condition.

[0111] Based on the same concept, the embodiments of the present application also provide an aircraft, which includes a fuselage, a wing, and a wing monitoring system provided by the embodiments of the present application. The wing monitoring system is used to perceive the vibration information and the deformation information of the wing, and predict the vibration information and the deformation variable of the wing.

[0112] The aircraft provided by the embodiments of the present application can accurately distinguish the vibration information and the deformation information of the wing in a complex dynamic environment by carrying the wing monitoring system provided by the embodiments of the present application, so as to realize high-precision and real-time perception of the vibration and deformation of the wing, and monitor and evaluate the health status of the wing.

[0113] Further, the TENG module and the EMG module in the sensor system in the wing monitoring system are integrated, and the PENG module is distributed on the wing, or the TENG module, the EMG module, and the PENG module are integrated and arranged on the wing.

[0114] Further, the sensor system can be installed at any position on the wing, and the maximum lift can be increased by 1.75%, and the maximum drag can be reduced by 0.8%. When the distance between the sensor system setting position and the wing leading edge point is equal to 30% of the wing chord length, the lift can be increased by 1.75%, and the drag can be reduced by 0.8%.

[0115] The above merely provides a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A wing monitoring system for an aircraft, characterized in that The application relates to a sensor system and a monitoring platform connected in communication, wherein the sensor system comprises a triboelectric nanogenerator (TENG) module, an electromagnetic generator (EMG) module and a piezoelectric nanogenerator (PENG) module installed on a wing of an aircraft for outputting a first electric signal representing wing vibration information, a second electric signal representing wing deformation information and a third electric signal representing wing deformation information, wherein the TENG module and the EMG module are used for outputting the first electric signal and the second electric signal respectively, and the PENG module is used for outputting the third electric signal; the monitoring platform is used for acquiring the first electric signal, the second electric signal and the third electric signal, and analyzing the first electric signal, the second electric signal and the third electric signal based on a pre-trained neural network model to predict the wing vibration information and the deformation amount; the pre-trained neural network model is generated by the following method: acquiring training sample data, wherein each training sample data comprises a first sample electric signal, a second sample electric signal, a third sample electric signal, sample vibration information and a sample deformation amount; performing the following operations on any training sample data until the training sample data are all processed or the prediction accuracy reaches a preset requirement to generate the neural network model: pre-processing the first sample electric signal, the second sample electric signal and the third sample electric signal to obtain three groups of preset length time series signals; processing the three groups of preset length time series signals by using a preset convolution kernel, extracting the time characteristics of each group of time series signals according to the target weight of each sample electric signal, and the time characteristics representing the relationship between the signal parameters and the time in the time series signal; splicing the time characteristics of the groups of time series signals to obtain spliced features, extracting dynamic evolution features from the spliced features through a bidirectional long short-term memory (LSTM) layer, and the dynamic evolution features representing the change relationship of the signal with time in the time series signal; outputting estimated vibration information and estimated deformation amount based on the dynamic evolution features, calculating a prediction error based on the estimated vibration information, the estimated deformation amount, the sample vibration information and the sample deformation amount through a preset loss function, and adjusting the target weight of each sample electric signal through a back propagation algorithm, wherein the preset loss function comprises a regularization loss term based on a vibration physical formula. The regularization loss term based on the vibration physical formula is shown in Formula One. The preset convolution kernel is three groups of different scale one-dimensional convolution kernels. The monitoring platform is further used for displaying the change curves of the first electric signal, the second electric signal and the third electric signal. The monitoring platform is further used for triggering an alarm prompt when the vibration information and / or the deformation amount meet a preset alarm condition. The TENG module comprises a first electrode, a second electrode, a permanent magnet, a plurality of support columns and a first load component. ​ ​ ​ ​ ​ 2. The system of claim 1, wherein, ​ wherein, is a regularization loss term, MSE is a mean squared error, represents a vibration acceleration in the estimated vibration information, represents a vibration frequency in the estimated vibration information, represents a vibration amplitude in the estimated vibration information.

3. The system of claim 1, wherein, ​ 4. The system of claim 1, wherein, ​ 5. The system of claim 1, wherein, ​ 6. The system of any one of claims 1-5, wherein, ​ The plurality of support columns are arranged between the first electrode and the second electrode, and a cavity is formed between the plurality of support columns, wherein the permanent magnet is arranged in the cavity, a first material is arranged on the surface of the permanent magnet, a second material is arranged on the surface of the first electrode opposite to the permanent magnet, and a second material or a third material is arranged on the surface of the second electrode opposite to the permanent magnet, the electronegativity of the first material is different from that of the second material, and the electronegativity of the first material is different from that of the third material, the first end of the first load assembly is connected to the first electrode, and the second end is connected to the second electrode.

7. The system of claim 6, wherein, The EMG module comprises a coil, a non-magnetic shell, and a second load assembly, the non-magnetic shell is arranged around the permanent magnet, the non-magnetic shell is arranged in the cavity, the coil is wound on the non-magnetic shell, the first end of the second load assembly is connected to one end of the coil, and the second end is connected to the other end of the coil.

8. The system of claim 1, wherein, The PENG module comprises a first piezoelectric unit, a second piezoelectric unit, and a differential circuit, the first piezoelectric unit is fixed to the surface of the wing through silica gel, the second piezoelectric unit is rigidly fixed to the first piezoelectric unit, and the input end of the differential circuit is connected to the first piezoelectric unit and the second piezoelectric unit respectively.

9. An aircraft, characterized in that The aircraft comprises a fuselage, a wing, and a wing monitoring system according to any one of claims 1-8, wherein, The wing monitoring system is used to sense the vibration information and deformation information of the wing, and predict the vibration information and deformation amount of the wing.

Images