A method of modal testing of an aircraft wing

Abstract

This invention discloses a method for modal testing of aircraft wings. It involves extracting mass concentration blocks and stiffness step blocks along the spanwise direction of the wing under test, and dividing it into characteristic segments with independent dynamic properties. Combining aircraft development requirements and flutter analysis frequency bands, the highest target modal order is determined. Then, based on the Nyquist space sampling theorem, the minimum number of measurement points is quantified and configured for each segment. The excitation signal and the response signal of each measurement point are acquired through sequential excitation and synchronous acquisition, ultimately reconstructing the target modal shape. This application eliminates sampling aliasing and modal loss caused by the shortening of the spatial wavelength of higher-order modes at the source. Simultaneously, the excitation method of instantaneous detachment after excitation eliminates the interference of the added mass of the traditional exciter on the natural frequencies of lightweight wings, significantly improving the identification completeness and data confidence of complex bending-torsional coupled modes and local high-frequency modes.

Description

Technical Field

[0001] This invention relates to the field of aviation technology, and in particular to a method for testing the modal characteristics of an aircraft wing. Background Technology

[0002] With the development of electric vertical takeoff and landing (EVTOL) aircraft, wings often utilize lightweight carbon fiber composite materials and carry distributed power components. These structures exhibit non-uniform mass and stiffness distribution along the spanwise direction, resulting in dense vibration modes and complex local deformations. During ground vibration testing, if the measurement point layout strategy fails to adequately consider the differences in the wing's physical characteristics, spatial sampling distortion can easily occur due to the shortening of higher-order mode wavelengths, making it difficult to accurately capture key modes and thus affecting the reliability of flutter analysis and flight safety assessment. Therefore, how to effectively arrange measurement points based on the wing's structural characteristics to improve the fidelity of higher-order mode identification has become an urgent technical problem to be solved. Summary of the Invention

[0003] This application provides a modal testing method for aircraft wings, which solves the technical problems in the prior art that lead to high-order modal space sampling distortion and inaccurate identification of bending-torsional coupling modes due to the failure to consider the differences in local physical characteristics of lightweight wings with non-uniform mass and stiffness distribution. It achieves the technical effect of high-fidelity extraction of high-order complex modal parameters and improved testing efficiency and data confidence by providing an adaptive quantitative measurement point layout based on the heterogeneous characteristics of the wing.

[0004] This application provides a method for modal testing of an aircraft wing, the method comprising: Based on the three-dimensional structural model of the wing under test, the mass concentration block and stiffness step block in the wing under test are determined. Based on the distribution positions of the mass concentration blocks and the stiffness step blocks on the wing under test, k characteristic segments arranged along the spanwise direction of the wing under test are determined; k is a positive integer. Based on the pre-set development requirements and flutter analysis frequency band of the aircraft on which the wing under test is located, the highest target mode order that needs to be extracted from the wing under test is determined; Based on the highest target modal order and the structural distribution characteristics of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined according to the Nyquist space sampling theorem. Excitation is applied to each target excitation position on the wing under test one by one. For each excitation process, the excitation signal in the current excitation process and the response signal of each target measurement point position in each feature segment are collected. The number of target measurement point positions in each feature segment is greater than or equal to the minimum number of measurement points corresponding to the feature segment, and satisfies the preset conditions of the intra-layer spatial topology distribution corresponding to the feature segment. Based on the locations of each target measurement point, the excitation signals and response signals of each excitation process, the target mode shape of the wing under test is constructed. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation.

[0005] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: This application embodiment extracts the mass concentration block and stiffness step block of the wing under test along the spanwise direction and divides it into characteristic segments with independent dynamic properties. Combining the aircraft development requirements and flutter analysis frequency band, the highest target mode order is determined. Then, based on the Nyquist space sampling law, the minimum number of measurement points is quantitatively configured for each segment. The excitation signal and the response signal of each measurement point are obtained by sequential excitation and synchronous acquisition. Finally, the target mode shape is reconstructed. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation. This application's embodiments transform the measurement point layout from an experience-based equidistant arrangement to a quantitative decision driven by both structural physical characteristics and sampling theorems. This fundamentally eliminates sampling aliasing and modal loss caused by the shortening of the spatial wavelength of higher-order vibration modes. At the same time, the excitation method of instantaneous detachment after excitation eliminates the interference of the added mass of the traditional exciter on the natural frequency of the lightweight wing. This ensures that the identified modal parameters truly reflect the dynamic characteristics of the wing in its free state, significantly improving the completeness and data confidence of the identification of complex bending-torsional coupling vibration modes and local high-frequency modes. This provides high-fidelity experimental data support for flutter boundary prediction and structural safety assessment. Attached Figure Description

[0006] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0007] Figure 1 A schematic flowchart illustrating a modal testing method for an aircraft wing provided in this application embodiment; Figure 2 This is a schematic diagram of a sparse distribution of measuring points provided in an embodiment of this application. Detailed Implementation

[0008] This application provides a modal testing method for aircraft wings, which solves the technical problems in the prior art that lead to high-order modal space sampling distortion and inaccurate identification of bending-torsional coupling modes due to the failure to consider the differences in local physical characteristics of lightweight wings with non-uniform mass and stiffness distribution. It achieves the technical effect of high-fidelity extraction of high-order complex modal parameters and improved testing efficiency and data confidence by providing an adaptive quantitative measurement point layout based on the heterogeneous characteristics of the wing.

[0009] The technical solution of this application embodiment is to solve the above-mentioned technical problems, and the general idea is as follows: This application embodiment extracts the mass concentration block and stiffness step block of the wing under test along the spanwise direction and divides it into characteristic segments with independent dynamic properties. Combining the aircraft development requirements and flutter analysis frequency band, the highest target mode order is determined. Then, based on the Nyquist space sampling law, the minimum number of measurement points is quantitatively configured for each segment. The excitation signal and the response signal of each measurement point are obtained by sequential excitation and synchronous acquisition. Finally, the target mode shape is reconstructed. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation. This application's embodiments transform the measurement point layout from an experience-based equidistant arrangement to a quantitative decision driven by both structural physical characteristics and sampling theorems. This fundamentally eliminates sampling aliasing and modal loss caused by the shortening of the spatial wavelength of higher-order vibration modes. At the same time, the excitation method of instantaneous detachment after excitation eliminates the interference of the added mass of the traditional exciter on the natural frequency of the lightweight wing. This ensures that the identified modal parameters truly reflect the dynamic characteristics of the wing in its free state, significantly improving the completeness and data confidence of the identification of complex bending-torsional coupling vibration modes and local high-frequency modes. This provides high-fidelity experimental data support for flutter boundary prediction and structural safety assessment.

[0010] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0011] First, it should be clarified that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0012] This application provides a method for modal testing of an aircraft wing, the method including steps S11-S16, which can be found in detail in reference to... Figure 1 As shown.

[0013] Step S11: Based on the three-dimensional structural model of the wing under test, determine the mass concentration block and stiffness step block in the wing under test. Step S12: Based on the distribution positions of the mass concentration blocks and the stiffness step blocks on the wing under test, determine k characteristic segments arranged along the spanwise direction of the wing under test; k is a positive integer. Step S13: Based on the preset development requirements and flutter analysis frequency band of the aircraft where the wing under test is located, determine the highest target mode order that needs to be extracted from the wing under test; Step S14: Based on the highest target modal order and the structural distribution characteristics of each feature segment, determine the minimum number of measurement points corresponding to each feature segment according to the Nyquist space sampling theorem. Step S15: Apply excitation to each target excitation position on the wing under test one by one. For each excitation process, collect the excitation signal in the current excitation process and the response signal of each target measurement point position in each feature segment. The number of target measurement point positions in each feature segment is greater than or equal to the minimum number of measurement points corresponding to the feature segment, and satisfies the preset conditions of the intra-layer spatial topology distribution corresponding to the feature segment. Step S16: Based on the location of each target measurement point, the excitation signal and response signal of each excitation process, construct the target mode shape of the wing under test. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation.

[0014] The aircraft wing modal testing method provided in this application can be executed by a computing control terminal with data processing and acquisition control capabilities, as well as a supporting excitation and sensing device. Specifically, it can be a testing system including a module with signal acquisition function, a module for triggering excitation, and modal analysis software.

[0015] Regarding step S11, based on the three-dimensional structural model of the wing under test, the mass concentration block and stiffness step block in the wing under test are determined.

[0016] The core of step S11 lies in extracting local physical feature regions closely related to the structural dynamics from the three-dimensional structural model of the wing under test. This embodiment does not rely on empirical estimation or uniformity assumptions, but rather automatically identifies concentrated mass locations (such as motors and pod mounting points) on the wing surface or interior where the mass density is significantly higher than the surrounding smooth areas, and stiffness transition boundaries where the cross-sectional bending stiffness changes abruptly (such as stiffener termination ends and bolt hole edges), through the analysis of three-dimensional geometric data. This process discretizes the continuous wing structure into a set of feature blocks with clear physical meaning, providing a quantitative spatial input basis for subsequent segmentation along the spanwise direction into independent dynamic feature layers and for implementing differentiated measurement point densification within each layer. This avoids the blind spots in modal testing caused by neglecting local heterogeneous features in traditional methods.

[0017] Specifically, based on the three-dimensional structural model of the wing under test, the mass concentration block and stiffness step block in the wing under test are determined, including: Step S111: Perform finite element preprocessing on the three-dimensional structural model of the wing under test to generate an initial element mesh; Step S112: Associate the initial cell mesh with the corresponding physical properties to obtain the target cell mesh; Step S113: Determine the linear mass density distribution curve and bending stiffness distribution curve along the spanwise direction of the wing under test based on the target cell mesh; Step S114: Determine the mass concentration area in the wing under test based on the linear mass density distribution curve; Step S115: Determine the stiffness step block in the wing under test based on the bending stiffness distribution curve.

[0018] Specifically, based on the three-dimensional structural model of the wing under test, the mass concentration blocks and stiffness step blocks are determined, which can be further refined into the following processing steps: First, a three-dimensional Cartesian coordinate system is established based on the fuselage symmetry plane or wing root of the distributed-powered aircraft, where the X-axis represents the flow direction, the Y-axis represents the spanwise direction, and the Z-axis represents the normal direction. The computer control terminal imports the three-dimensional CAD geometric model of the wing and the mounted arm under test through a general interface. For physical verification aircraft without drawings, the wing shell is reverse-engineered and reconstructed using three-dimensional laser scanning or industrial photogrammetry technology to obtain point cloud data and fit it into a three-dimensional geometric envelope map.

[0019] Next, the obtained three-dimensional geometric model or envelope diagram is imported into finite element preprocessing software (such as Hypermesh) for mesh generation, generating preliminary shell element and volume element meshes. Based on the aircraft bill of materials or design drawings, the actual physical properties such as material density, elastic modulus and composite material skin ply thickness are assigned to the corresponding mesh elements. For complex interior components that cannot be modeled in detail, they are equivalent to concentrated mass points applied to the corresponding installation nodes.

[0020] Based on this, the linear mass density distribution curve and bending stiffness distribution curve per unit length are calculated along the wing span: the former shows a significant peak at physical mounting locations such as motors, rotor pods, heavy servo motors, or in-wing batteries, thus automatically identifying local mass concentration blocks; the latter shows discontinuous jumps at the junction of the mechanical connection between the arm and the wing, the reinforced wing ribs, the thickened section of the main spars, or the local skin thickening area, thus locking in the stiffness step block.

[0021] Finally, the start and end coordinate parameters of these mass peaks and stiffness jump points in the spanwise direction are extracted. The complex heterogeneous region where mass concentration and stiffness step are coupled is accurately delineated on the three-dimensional geometric model, and its spatial boundary data is output. This provides a precise physical range basis for subsequent segmentation of characteristic layers along the spanwise direction and implementation of differentiated measurement point densification.

[0022] Taking the left wing of a certain type of distributed-powered electric vertical takeoff and landing (eVTOL) aircraft as an example, the wing has a span of approximately 4 meters, and a main lift motor assembly weighing approximately 12 kilograms is mounted 1.8 to 2.5 meters from the wing root. First, the STEP format 3D CAD model of the wing is imported into the computer control terminal, and a coordinate system is established with the intersection of the longitudinal symmetry plane of the fuselage and the leading edge of the wing as the origin. The X-axis is in the direction of airflow, the Y-axis points towards the wingtip along the span, and the Z-axis is perpendicular to the wing skin and upward. Subsequently, the model was imported into Hypermesh software for shell element mesh generation, producing approximately 25,000 quadrilateral elements. Based on the bill of materials, the carbon fiber skin region was assigned a density of 1.6 g / cm³ and an elastic modulus of 70 GPa, while the aluminum alloy adapter plate region of the motor pod was assigned a density of 2.8 g / cm³ and an elastic modulus of 72 GPa. The motor body was equivalently represented as a CONM2 lumped mass element located at its center of gravity projection point. After completing the physical property association, the algorithm calculated the spanwise mass density curve and bending stiffness curve along the Y-axis with a step size of 10 mm. On the linear mass density curve, a sharp peak appeared in the Y-coordinate range of 1.8 m to 2.0 m, with the linear density jumping from approximately 3 kg / m to 18 kg / m; this area was automatically marked as the first mass concentration block. A second mass peak appeared in the Y-coordinate range of 2.3 m to 2.5 m due to the fixing lugs at the rear of the motor. On the bending stiffness distribution curve, at the Y-coordinate of 1.75 meters, due to the initial boundary of the reinforcing rib of the motor front mounting bracket, the bending stiffness of the section increases from approximately 2.5 × 10⁻⁶ m. 4 The N·m² suddenly jumped to 8.7 × 10⁻⁶. 4 The stiffness increases significantly at a point N·m², resulting in a noticeable stiffness step. Similarly, at 2.55 meters on the Y-coordinate, the termination of the reinforced structure causes the stiffness to drop sharply back to the baseline level. Based on this, the system outputs the spanwise start and end coordinate ranges of the aforementioned mass concentration and stiffness step blocks, and marks the motor-mounted coupling heterogeneous region in the 3D model with a red highlight circle. This provides precise geometric boundary input for subsequently dividing the wing into a root reinforcement layer, a smooth transition layer, a power component mounting layer, and a wingtip free layer.

[0023] Regarding step S12, based on the distribution positions of the mass concentration block and the stiffness step block on the wing under test, k characteristic segments arranged along the span of the wing under test are determined; k is a positive integer.

[0024] Step S12 involves using the spanwise coordinate boundaries of the mass concentration blocks and stiffness step blocks on the wing surface, obtained from previous steps, as the basis for partitioning. This divides the originally continuously varying wing along the spanwise direction into k independent feature segments with significant dynamic differences. This process is not a simple geometric division, but rather uses the mass peak position and stiffness abrupt change boundary as the separating surface between adjacent segments. This ensures that each segment maintains relative uniformity or continuity in mass distribution and stiffness variation trends, while exhibiting significant dynamic property jumps between segments. Thus, the complex wing, originally exhibiting a non-uniform heterogeneous distribution along the spanwise direction, is transformed into a discretized model composed of several sequentially adjacent feature segments. Each segment corresponds to a typical local structural feature, providing a clear spatial operation unit and logical basis for subsequently determining the spatial sampling density based on the physical complexity of each segment and implementing targeted measurement point densification strategies.

[0025] Specifically, based on the distribution positions of the mass concentration blocks and the stiffness step blocks on the wing under test, k characteristic segments arranged along the spanwise direction of the wing under test are determined, including: Step S121: Based on the distribution positions of the mass concentration block and the stiffness step block on the wing under test, determine the mass density gradient function and bending stiffness gradient function of the wing under test along the spanwise direction. Step S122: The sum of the number of extreme points in the mass density gradient function and the number of jump points in the bending stiffness gradient function is determined as the number of characteristic segments arranged along the span of the wing under test.

[0026] Given that the eVTOL wing exhibits non-uniform heterogeneous characteristics along its span, this embodiment of the application abandons global equidistant segmentation and divides the test wing along its span into k levels with independent dynamic characteristics (where k≥3).

[0027] In practice, the starting and ending coordinates of the mass concentration block and stiffness step block extracted in the previous steps in the spanwise direction are first used to construct the linear mass density gradient function and bending stiffness gradient function of the wing along the Y-axis. The former is obtained by taking the first derivative of the linear mass density distribution curve, and the extreme point is the section position where the mass changes most drastically along the spanwise direction. The latter is obtained by performing gradient calculation on the bending stiffness curve of the section and detecting the step position, and the jump point corresponds to the connection boundary where the stiffness changes abruptly.

[0028] Subsequently, the number of extreme points in the mass density gradient function and the number of jump points in the bending stiffness gradient function are summed, and the total number obtained is the number of characteristic segments k to be divided along the span. In this embodiment, the setting of k≥3 is to establish the simplest dynamic identification chain of "constrained end-heterogeneous region-free end", where the specific value of k is objectively determined by the statistical results of the above extreme points and jump points. When there are multiple power component mounting areas in the wing, the gradient function will present multiple pairs of extreme points and jump points. By increasing the value of k, decoupling sampling of each local nonlinear feature is achieved, thereby completely avoiding spatial aliasing error at the algorithm level.

[0029] In dynamic modeling, k=3 is the smallest topological unit that abstracts the heterogeneous wing of eVTOL from the "traditional wing": The first layer is the fixed end of the fuselage root, which is the stiffness abrupt change point and the strongest constraint area of ​​the structure. Its modal characteristics are significantly affected by the fuselage coupling. Without this layer, the subsequent mode shapes cannot obtain absolute reference coordinates. The second layer is the heterogeneous dynamics region, which is the core area of ​​concern for eVTOL. Due to the mounting of motors and arms, the mass and stiffness undergo a step change here. Therefore, this region must be independently layered to properly apply the encryption algorithm. >1.0; The third layer is the wingtip free end, which is the region with the largest structural displacement amplitude and the most sensitive modal recognition signal-to-noise ratio.

[0030] Preferably, the k layers, from the wing root to the wingtip, include at least the following: a root reinforcement layer fixed to the fuselage, a smooth mid-section transition layer without power attachment, a power assembly mounting layer with distributed motors, and a freely suspended tip free end layer.

[0031] Taking the left wing of a certain type of eVTOL with a wingspan of 4 meters as an example: along the mass density curve calculated along the span, two obvious mass peaks were detected at Y=1.8m and Y=2.5m, corresponding to the two extreme points of the mass density gradient function; on the bending stiffness curve, two stiffness jumps occurred at Y=1.75m and Y=2.55m, corresponding to the two jump points of the bending stiffness gradient function. The sum of the number of extreme points (2) and the number of jump points (2) is 4, so k=4 is initially determined. Further comparison with the requirements of the dynamic identification chain: Y=0 to 0.5m is the root reinforcement layer fixed to the fuselage, Y=0.5m to 1.7m is the smooth mid-section transition layer without load, Y=1.7m to 2.6m is the power component installation layer with the motor, and Y=2.6m to 4.0m is the wingtip free end layer, a total of 4 characteristic layers, consistent with the calculation results. If the other wing also carries additional concentrated mass such as auxiliary servos, an extreme point will be added to the gradient function, and the value of k will increase accordingly. This ensures that each local heterogeneous feature is independently divided into a segment, thereby applying denser sampling to the area when subsequent measurement points are deployed.

[0032] Regarding step S13, based on the preset development requirements of the aircraft where the wing under test is located and the flutter analysis frequency band, the highest target mode order that needs to be extracted from the wing under test is determined.

[0033] The core of step S13 lies in defining a clear order extraction range for subsequent modal tests, starting from the top-level design specifications and aeroelastic safety boundaries of the aircraft. Based on the minimum coverage frequencies implicit in the development requirements of the aircraft containing the wing under test (e.g., 2 to 3 times the required bandwidth for flight control stabilization, the characteristic frequency band of the motor for vibration and noise control, and the mandatory mode order specified in airworthiness certification clauses) and the upper limit of the safety margin required for flutter analysis frequency bands, a continuous target test frequency range covering both is first comprehensively determined. Then, a finite element model is used to pre-analyze this frequency range, counting the number of inherent frequency characteristic values ​​contained within this interval. This number corresponds to the highest target mode order that must be extracted with high fidelity. Through this process, the test system does not need to blindly scan all possible modes within the frequency range during implementation. Instead, it precisely focuses algorithm resources, sensor sampling density, and excitation energy on the key mode set defined by this order, thereby significantly improving test efficiency and the data confidence of the final mode shape reconstruction while ensuring the integrity of the flutter risk assessment.

[0034] Specifically, the pre-set development requirements include at least flight control bandwidth requirements, vibration and noise control requirements, and airworthiness certification requirements.

[0035] Flight control bandwidth requirements stem from engineering constraints in the field of aerodynamic servo-elasticity, referring to the upper limit of sensor sampling and actuator response frequencies relied upon by the flight control system for real-time stabilization control to maintain aircraft attitude stability. On eVTOL aircraft, the flight control computer continuously senses the airframe angular rate and acceleration via an inertial measurement unit. If the wing structure's inherent vibration frequency falls within this effective bandwidth (e.g., the flight control loop cutoff frequency is 15 Hz, while the wing's first-order bending mode happens to be near 12 Hz), the flexible deformation signal will be misinterpreted by the flight control as rigid body motion, inducing unstable coupling between the control system and structural vibrations, which can lead to flight loss of control in severe cases. Therefore, flight control bandwidth requirements mandate that modal testing must extend the effective operating frequency band of the flight control system upwards by 2 to 3 times as a minimum testing range to ensure that all possible structural modes within this range are identified and recorded.

[0036] Vibration and noise control requirements focus on the acoustic and vibration environment within the passenger cabin and the limitations imposed by the durability of onboard equipment on the structural frequency layout. The eVTOL propulsion system uses multiple high-power motors to drive propellers, generating lift and thrust. During operation, the motor rotors output a fundamental frequency and its integer multiples of harmonics that strictly correspond to the rotational speed. If components such as the wing skin, floor support structure, or battery mounts exhibit inherent modes within the motor's operating speed range, local resonance will occur. This not only produces annoying single-frequency humming noise, reducing ride quality, but may also cause fatigue cracks in connection joints or malfunctions in avionics. Therefore, the reliability design department calculates the corresponding upper limit of the excitation frequency based on the motor's maximum continuous operating speed (e.g., a maximum speed of 1800 rpm corresponds to a fundamental frequency of 30 Hz, and the six-blade pass frequency is 180 Hz). Based on this, clear verification requirements are proposed for modal testing, demanding that the tests fully cover this upper limit frequency.

[0037] Airworthiness certification requirements are the legally mandated criteria that civil aircraft must meet before entering commercial operation. These requirements are issued by national civil aviation authorities in the form of specific conditions or advisory circulars. For novel configurations like eVTOLs, current regulations do not fully cover all their technical characteristics. Certification authorities typically refer to existing airworthiness standards for fixed-wing and rotary-wing aircraft, combined with the conclusions of specific technical research, to formulate applicable clauses. Ground vibration testing, as crucial compliance evidence for verifying the accuracy of structural dynamics models and confirming the reliability of flutter boundary calculations, is listed as a mandatory closure item in the certification plan. For example, the certification basis for a certain eVTOL model might explicitly stipulate that the applicant must obtain the first few major symmetric and antisymmetric modal parameters through full-aircraft ground resonance testing and verify their correlation with the flutter analysis model. This clause is only approved and closed if the model predictions and test results have a frequency deviation of less than 3% and a modal confidence criterion higher than 0.9.

[0038] Furthermore, based on the pre-defined development requirements and flutter analysis frequency band of the aircraft containing the wing under test, the highest target mode order to be extracted from the wing under test is determined, including: Step S131: Based on the preset development requirements of the aircraft containing the wing under test and the minimum test frequency corresponding to each flutter analysis frequency band, determine the target test frequency range of the aircraft; the target test frequency range covers the minimum test frequency corresponding to each preset development requirements and flutter analysis frequency band. Step S132: Perform finite element analysis on the target test frequency range to determine the number of natural frequency eigenvalues ​​of the target test frequency range, and determine the number of natural frequency eigenvalues ​​as the highest target modal order.

[0039] The decision logic of step S131 is essentially a "index coverage" process: First, the minimum test frequency boundaries required by each of the multiple technical sources are collected in parallel. For example, the flight control department proposes the upper limit of the frequency that must be covered based on the flight control bandwidth requirements (e.g., requiring testing up to 40 Hz to avoid the risk of coupling in the flight control loop). The power and noise control department proposes its own verification frequency band based on the maximum operating speed of the motor and the passing frequency of its blades (e.g., requiring testing up to 80 Hz to confirm that the wing does not resonate under the main harmonics of the motor). The aeroelasticity department proposes a safety monitoring range based on the unstable range predicted by flutter analysis (e.g., predicting that the flutter risk is concentrated around 60 Hz, and requiring it to be extended upward to 1.5 times that frequency, i.e., 90 Hz, to cover all potential coupled vibration modes). Flutter is the most dangerous aeroelastic phenomenon in aircraft, representing an inviolable safety red line and the core purpose of modal testing to calibrate flutter calculation models. If flutter analysis predicts instability near 30 Hz, it is essential to ensure that subsequent modal extraction orders fully cover this frequency point and extend to at least 1.5 times the expected flutter frequency to prevent omission of relevant higher-order bending-torsional coupling modes. Subsequently, the computer control terminal selects the maximum value from the aforementioned parallel inputs (80 Hz as required by the power department in this example), thereby uniquely determining a continuous and closed target test frequency range, for example, defined as 0 to 80 Hz. This range fully covers all minimum test frequency specifications proposed by various stakeholders.

[0040] Step S132 maps the frequency range obtained in step S131 into executable operation instructions: During the preliminary research phase, a modal analysis is performed using the established finite element model of the wing under test. The solution frequency range is set to be exactly the aforementioned target test frequency range (0 to 80 Hz). Within this limit, the software sequentially outputs all structural eigenvalues ​​that meet the convergence conditions. In engineering context, eigenvalues ​​correspond to each order of elastic natural frequencies, and each eigenvalue is uniquely associated with a first-order mode shape. The system counts the total number of eigenvalues ​​contained in the natural frequency list arranged from low to high. For example, if 12 valid eigenvalues ​​are calculated within this range, it means that the wing has 12 natural modes within the 0 to 80 Hz interval. This number is directly extracted and assigned to the highest target mode order. That is, setting =12. At this point, the multiple top-level indicators from the research and development requirements and flutter analysis have been quantitatively transformed into a precise integer instruction. All subsequent operations regarding spatial sampling density calculation, sensor quantity allocation, and signal processing bandwidth setting are strictly centered around these 12 modes that must be extracted with high fidelity.

[0041] Taking a certain type of eVTOL demonstrator as an example, the flight control unit requires coverage up to 40 Hz, the power unit requires coverage up to 80 Hz (six times the motor's passing frequency), and the aeroelasticity unit predicts the flutter sensitive area to be around 55 Hz and requires coverage up to 1.5 times that, i.e., 82.5 Hz. Ultimately, the maximum value of 80 Hz was selected as the target test frequency range. After performing finite element analysis on this range, the eigenvalue list obtained is as follows: 1st order 12.1 Hz, 2nd order 22.3 Hz, 3rd order 31.8 Hz, 4th order 42.5 Hz, 5th order 51.2 Hz, 6th order 58.7 Hz, 7th order 63.4 Hz, 8th order 67.9 Hz, 9th order 72.1 Hz, 10th order 75.3 Hz, 11th order 77.8 Hz, 12th order 79.6 Hz. There are 12 eigenvalues. The value was determined to be 12. Based on this, the test system anchored the sensor deployment density, excitation energy spectrum shape, and algorithm pole search order to these 12 modes in subsequent steps, ensuring that test resources were configured efficiently while meeting all compliance requirements.

[0042] Regarding step S14, based on the highest target modal order and the structural distribution characteristics of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined according to the Nyquist space sampling theorem.

[0043] Step S14 is the core quantitative step of converting the theoretically highest target modal order into the physical measurement point layout. This is after the highest target modal order has been obtained. Based on the k characteristic segments divided along the span, this step introduces the Nyquist spatial sampling theorem to differentiate the spatial oversampling requirements of each segment according to the structural distribution characteristics reflected by the mass concentration and stiffness gradient abrupt change within each segment: For transitional sections with gradual changes in mass and stiffness, only the theoretical minimum sampling density needs to be met. However, for sections containing local heterogeneous features such as dynamic loads and stiffener termination lines, since the spatial wavelength of higher-order modes is severely compressed and the distribution of nodal lines is extremely unstable, an oversampling factor greater than 1 is forcibly assigned, and a leakage prevention constant related to sensor channel redundancy is superimposed. Finally, the minimum number of measurement points for this section is generated by rounding up.

[0044] Through this process, the measurement point layout, which originally relied on engineers' experience for rough estimation, has been transformed into a set of quantitative decision-making criteria driven entirely by structural physical characteristics and sampling theorems. This ensures that in the subsequent response acquisition stage, each characteristic segment can capture the complete local deformation waveform at a sampling frequency higher than that required for the shortest wavelength of the mode. This eliminates spatial aliasing caused by undersampling from a mathematical perspective, and provides sufficient sensor spatial degrees of freedom for distortion-free reconstruction of complex high-order vibration modes such as bending-torsional coupling.

[0045] Specifically, based on the highest target modal order and the structural distribution characteristics of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined according to the Nyquist space sampling theorem, including: Step S141: Based on the structural distribution characteristics of each feature segment and the preset correlation relationship, determine the spatial oversampling ratio corresponding to each feature segment; the preset correlation relationship characterizes the correspondence between the structural distribution characteristics and the spatial oversampling ratio; the preset correlation relationship is determined based on the Nyquist spatial sampling theorem. Step S142: Based on the highest target modal order, the preset leakage redundancy constant, and the spatial oversampling ratio of each feature segment, determine the minimum number of measurement points corresponding to each feature segment. Refer to the preset formula for details.

[0046] The preset formula is:

[0047] in, Indicates the first The minimum number of measurement points corresponding to each feature layer segment Indicates to Round up. Indicates the first Spatial oversampling ratio corresponding to each feature segment This indicates the highest target modal order. This represents the preset leakage prevention redundancy constant.

[0048] In the specific implementation process, the computer control terminal first reads the highest target modal order input in the aforementioned steps from memory. Simultaneously extract the total number k of the divided wing feature segments and the number of each segment. ( The three-dimensional spatial boundary coordinates (i = 1, 2, ..., k) are used to complete the data preparation before execution. Subsequently, based on the structural distribution characteristics of each layer and a preset correlation, the terminal automatically determines the corresponding spatial oversampling ratio for each layer. .

[0049] This pre-defined correlation is based on the Nyquist space sampling theorem and establishes a mapping logic for the non-uniform heterogeneous characteristics of the eVTOL wing: for an ideal uniform cantilever beam, extract the preceding... In theory, only one mode is needed. A single spatial sampling point is sufficient to meet the minimum requirements of the sampling theorem; however, real eVTOL wings exhibit significant mass and stiffness differences along the span. The smooth transition section and the root reinforcement layer, due to their gradual structural wavelength changes, are deemed not to require excessive oversampling, and the terminal automatically assigns weights based on the mapping table. =1.0. However, in sections containing stiffness steps and mass concentration zones with mounted power components, the spatial wavelength of local high-frequency vibrations is severely compressed, making them highly susceptible to spatial undersampling and aliasing distortion under normal sampling densities. The terminal then automatically assigns weights based on the peak mass gradient or stiffness jump amplitude in that region. >1.0, for example, for heavy-duty motor mounting areas, can be taken as... =1.5, thereby forcibly increasing the local spatial resolution of this layer.

[0050] Based on this, considering that there must exist nodal lines with zero amplitude in the mode shape, if the sensor measurement point happens to be located on one of them, it will cause the energy of that mode to be completely lost. Therefore, the terminal program further introduces a preset leakage prevention redundancy constant. Typically, the redundancy of the acquisition system channels is used to determine the redundancy. Set it to a positive integer 1 or 2 to probabilistically avoid this kind of coincidental failure.

[0051] Complete each layer empowerment and After the value is set, the terminal starts a loop calculation program for each segment. The preset formulas are substituted sequentially to perform quantization calculations. First, the original theoretical values ​​are calculated. This value comprehensively reflects the basic requirements of the target modal order for spatial sampling, the mandatory requirements of local structural complexity for oversampling, and the compensation of sensor redundancy for nodal line blind spots. Since the actual number of deployable accelerometers must be a physical integer, the terminal... Force the floor function to be executed, i.e. Thus, the first The minimum number of measurement points that cannot be compromised in physical space for each feature segment.

[0052] It should be noted that the above point calculation formula has an algorithmic substitution form: besides relying on discrete layer segment division and multiplying by a coefficient, In addition, the sensor distribution density can be directly controlled by a continuous mass and stiffness gradient function, so that the distribution density changes proportionally to the concentration of structural mass or the degree of stiffness step at the location. This alternative solution also falls within the protection scope of this method.

[0053] Finally, after all k segments have completed their iterative calculations, the computer control terminal generates a one-dimensional spatial quantization measurement point allocation array [N_1, N_2, ..., N_k], and outputs it as a hard indicator to the monitor of the front-end testers. This clearly instructs the testers to accurately arrange the corresponding number of measurement points within the specified physical layer area, thereby completely closing the spatial anti-aliasing logic in the front-end point placement stage.

[0054] It is important to note that these measurement points can be measured using contact accelerometers, non-contact 3D laser Doppler vibrometers, or high-speed binocular stereo vision cameras. In other words, besides using piezoelectric accelerometers bonded to the wing, spatial scanning tests can be performed using non-contact 3D laser Doppler vibrometers (3D-LDV), or target tracking can be achieved using a high-speed binocular stereo vision camera combined with digital image correlation (DIC) technology.

[0055] If contact-type accelerometers are used for measurement, then the number of measurement points required for each wing level, as calculated above, will be... The coordinates of the geometric mesh intersections or internal rib stress points of the wing finite element model are used as physical measurement points. Miniature piezoelectric accelerometers (rigidly attached with aviation beeswax or adhesive) are then deployed at the rigid point locations of the arm determined in the aforementioned steps to construct a distributed sensing network.

[0056] Specifically, based on the preset conditions of the intra-layer spatial topology distribution corresponding to each feature segment, the target measurement point positions for deploying accelerometers are determined in each feature segment, and accelerometers are rigidly attached to each target measurement point using aviation beeswax or adhesive. For each excitation process, the excitation signal during the current excitation process and the response signal at each target measurement point position in each feature segment are collected by the accelerometers at each target measurement point position.

[0057] Taking the left wing of a certain type of eVTOL aircraft with a wingspan of 4 meters as an example: after completing the characteristic segment division, the wing is divided from the wing root to the wingtip into k=4 segments: wing root reinforcement layer (Y=0-0.5m), smooth mid-section transition layer (Y=0.5-1.7m), power assembly mounting layer (Y=1.7-2.6m), and wingtip free end layer (Y=2.6-4.0m). The highest target modal order has been determined based on development requirements and flutter analysis. =12, and the leakage redundancy constant β is set to 1. After the terminal reads the boundaries of each segment, weights are assigned to the root reinforcement layer and the smooth middle section transition layer respectively. = =1.0; Due to the presence of a 12 kg main motor load and two stiffness jumps in the power component mounting layer, the weighting is adjusted accordingly. =1.5; Although there is no concentrated mass in the wingtip free end layer, it is still weighted due to its sensitivity to free end amplitude. =1.2. Substitute into the formula to calculate the original theoretical value: =1.0×12+1=13, =1.0×12+1=13, =1.5×12+1=19, =1.2×12+1=15.4; After rounding up, the minimum number of measurement points for each layer is as follows: =13, =13, =19, =16. The terminal finally outputs the distribution array [13, 13, 19, 16] on the display. Based on this, the testers densely deployed 19 sensors in the power component mounting layer, 16 in the wingtip layer, and 13 each in the midsection and root layer to ensure that the spatial wavelengths of higher-order coupled vibration modes are also completely captured in the mass concentration area.

[0058] For example, such as Figure 2 As shown, the number of measurement points in regions A and B (power component mounting layer - complex heterogeneous region) is relatively dense, while the number of measurement points in other regions (wingtip free end layer and root reinforcement layer - uniform smooth region) is relatively sparse.

[0059] Furthermore, when flutter analysis or higher-order torsional mode identification places higher demands on the signal-to-noise ratio of the wingtip region, the measurement point density of the wingtip free-end layer can be actively increased based on the calculation using the preset formula. For example, several measurement points can be added at the spanwise ends of the wingtip leading and trailing edges to bring the sampling density closer to that of the power assembly mounting layer, thereby further enhancing the spatial capture capability of local high-frequency vibration modes at the point of maximum amplitude at the free end.

[0060] Regarding step S15, excitation is applied to each target excitation position on the wing under test one by one. For each excitation process, the excitation signal in the current excitation process and the response signal of each target measurement point position in each feature segment are collected. The number of target measurement point positions in each feature segment is greater than or equal to the minimum number of measurement points corresponding to the feature segment, and satisfies the preset conditions of the intra-layer spatial topology distribution corresponding to the feature segment.

[0061] Step S15 aims to transform the theoretical measurement point layout and excitation strategy determined by the previous calculations into executable physical test operations, which is a key execution link connecting algorithm planning and data acquisition.

[0062] During the testing phase, several typical target excitation locations are first selected based on the structural characteristics of the wing. These typically cover the trailing edge of the wing to excite the global bending and torsional modes, as well as the high-stiffness points at the far end of the power component mounting layer to excite the local high-frequency vibration modes. A point-by-point sequential excitation method is then used to apply transient pulses to each excitation location in turn.

[0063] During each excitation process, the data acquisition system uses the rising edge of the excitation signal (such as a force signal) as the hardware trigger reference, and synchronously records the force input signal corresponding to the impact, as well as the acceleration response signals of all target measurement points located in each feature layer.

[0064] The actual number of target measurement points deployed within each feature segment strictly adheres to the physical constraints required by that segment: on the one hand, the number of measurement points must not be less than the minimum number of measurement points calculated using the anti-aliasing formula in the aforementioned steps. This ensures that the layer has the minimum sensing density that satisfies the Nyquist space sampling theorem at the highest target modal order; on the other hand, the specific spatial distribution of each measuring point within the layer must meet the preset conditions of the spatial topology distribution within the layer.

[0065] The preset conditions for the spatial topology distribution within the layer include: avoiding the target measurement point location being in the flexible central region of the wing skin; maximizing the envelope of adjacent target measurement point locations in the spanwise and chordwise directions of the wing under test; and ensuring that the density of target measurement point locations in the feature layer segment containing the power mount is greater than the density of target measurement point locations in the feature layer segment not containing the power mount.

[0066] The pre-defined spatial topology distribution conditions within a layer are a set of spatial arrangement rules that testers must follow when deploying accelerometers within characteristic layer segments. Their purpose is to transform the theoretically required number of measurement points satisfying the Nyquist sampling theorem into a physically sound sensor layout with high signal-to-noise ratio, strong anti-interference capability, and favorable mode shape reconstruction. The computer control terminal outputs the... Minimum number of measurement points required for each feature segment Afterwards, the testers placed this in the physical space. When installing sensors, the following specific constraints must be implemented one by one, and a blind deployment method with uniform and equidistant spacing is not allowed.

[0067] First, implement structural hard-point constraints to avoid skin effects. Sensors must not be placed in the flexible central region of the wing skin. When the wing surface, especially skin made of carbon fiber laminate, is subjected to excited vibration, the local bulging modes of the shell introduce high-frequency noise unrelated to the deformation of the main load-bearing structure. If sensors are attached to unsupported areas of the skin, the collected response signals will be severely contaminated by these spurious high-frequency components. Therefore, within the layer... Each measuring point must be forcibly attached to the projection point of the main load-bearing frame inside the wing onto the skin surface, i.e., the "structural hard point". Preferred locations include the geometric intersection of the wing's leading-edge main spars and ribs, the intersection of the trailing-edge secondary spars and ribs, and the flange edge of the power assembly metal mounting bracket. These locations directly and rigidly couple the sensor to the macroscopic load-bearing path of the wing, which can physically filter out worthless local skin bulging noise and ensure that the acquired signal faithfully reflects the wing's dominant bending and torsional vibration modes.

[0068] Secondly, a strategy of maximizing spanwise and chordwise envelopes is implemented to enhance the sensitivity in capturing torsional modes. In this... When assigning measurement points to candidate hard point locations within a segment, the distance between the measurement points along the wing's chord direction (X-axis) must be maximized. Specifically, measurement points should be arranged in pairs or staggered at corresponding positions on the leading-edge main spars and trailing-edge secondary spars of the airfoil section, forming a "leading-tailing" span envelope pattern. This spatial distribution can detect the torsional deformation of the wing around its spanwise axis with extremely high sensitivity, i.e., the torsional component in the bending-torsional coupled mode. If the measurement points are too concentrated in the chord direction, the relative torsional displacement of the upper and lower surfaces of the wing or the leading and trailing edges will be submerged in the bending component in the signal due to the short lever arm, causing the subsequent modal identification algorithm to be unable to decouple a pure torsional mode shape.

[0069] Secondly, the density contrast of the measuring points is adjusted according to the density stratification principle. In the characteristic layer section containing dynamic loads, as mentioned above, the formula has already been used to adjust the density contrast. The coefficient forcibly increases the minimum number of measurement points required. During actual deployment, the density of target measurement points in this section should be significantly higher than that in smooth transition sections or free wingtip sections without powered loads. The high-density deployment area allows sensors to be densely arranged over a short distance, thus fully sampling the high-order spatial waveforms that are severely compressed at the concentrated mass and stiffness step. In the boom section without powered loads or the transition section that only serves to transmit force, only a small number of rigid measurement points need to be arranged at the root and end to define the overall torsional and bending trends of the section, thereby achieving a clear system-level topology deployment.

[0070] Finally, the regular incremental numbering constraint of coordinate mapping must be strictly enforced to prevent spatial coordinate errors during mode shape reconstruction. The measurement points and their corresponding data acquisition channel numbers within all wing segments must be arranged according to the regularity of spatial geometry. For example, strictly following the sequence from the spanwise root on the left side of the wing, through the central fuselage, to the right wingtip, with the hardware channel numbers of each measurement point increasing sequentially from smallest to largest (i.e., P1, P2, P3… consecutively). For paired measurement points at the same spanwise position but with chordal orientations at the leading and trailing edges, a consistent incremental numbering rule should also be established (e.g., leading edge then trailing edge). This regular numbering order ensures that in subsequent steps, the row vector of the frequency response function matrix indexed by the sensor number can be uniquely and correctly mapped to the corresponding node coordinates in the 3D geometric model, avoiding incorrect assignment of the residue data of the wing root measurement points to the wingtip mesh due to channel mapping misalignment, which would result in non-physical spatial distortions in the reconstructed mode shape animation.

[0071] Taking the power assembly mounting layer of a certain type of eVTOL wing as an example: the minimum number of measuring points allocated to this layer. =19. The testers first consulted the wing structure drawings and identified three intersection points between the leading-edge main spars and the three wing ribs within the spanwise range of this section, as well as three intersection points between the trailing-edge secondary spars and the same three wing ribs. They also identified two hard points along the flange edges of two motor mounting brackets, obtaining a total of eight candidate structural hard points that met the requirements. Since 19 is greater than 8, some measuring points needed to be fine-tuned in nearby positions or supplemented using adjacent ribs. During deployment, the testers deliberately placed three sensors at the intersection of the leading-edge wing ribs and three at the corresponding positions on the trailing-edge side, forming three sets of leading-edge and trailing-edge span pairs. Additional points were placed at the bracket flanges to ensure sensor coverage in both the leading and trailing-edge directions of this section. Simultaneously, the average spacing between measuring points in this section was approximately 0.15 meters, far less than the 0.3-meter spacing in the adjacent smooth transition section, reflecting density stratification. All measuring points are numbered sequentially from P21 to P39, following the rule from wing root to wingtip, and from the leading edge to the trailing edge. This thoroughly solidifies the spatial topology logic during the physical installation phase, establishing a precise coordinate mapping foundation for subsequent high-fidelity bending-torsional coupling mode reconstruction.

[0072] Through the above operations, the vibration energy generated by each strike is transmitted to each characteristic segment of the wing via different paths, and is completely captured by a sensor network with pre-optimized density in a spatial anti-aliasing manner. This provides spatiotemporally synchronized, signal-to-noise ratio-sufficient, and spatially distorted raw data support for the subsequent construction of the frequency response function matrix and high-fidelity identification of higher-order bending and torsional coupling modes.

[0073] If contact-type accelerometers are used for measurement, then the number of measurement points required for each wing level, as calculated above, will be... The coordinates of the geometric mesh intersections or internal rib stress points of the wing finite element model are used as physical measurement points. Miniature piezoelectric accelerometers (rigidly attached with aviation beeswax or adhesive) are then deployed at the rigid point locations of the arm determined in the aforementioned steps to construct a distributed sensing network.

[0074] Specifically, based on the preset conditions of the intra-layer spatial topology distribution corresponding to each feature segment, the target measurement point positions for deploying accelerometers are determined in each feature segment, and accelerometers are rigidly attached to each target measurement point using aviation beeswax or adhesive. For each excitation process, the excitation signal during the current excitation process and the response signal at each target measurement point position in each feature segment are collected by the accelerometers at each target measurement point position.

[0075] Furthermore, the target excitation location includes at least one of the following locations: the trailing edge of the wing and the high-stiffness point at the far end of the power assembly mounting layer.

[0076] The selection of the target excitation location is not arbitrary, but rather an optimal choice based on a deep understanding of the excitation mechanisms of the global and local modes of the wing. The trailing edge of the wing, due to its thinner skin, distance from the main load-bearing beam, and weaker boundary constraints, allows for a larger normal travel when struck, enabling energy to be smoothly transferred along the spanwise direction to the entire wing. This is particularly suitable for exciting low-frequency overall bending and torsional modes. Meanwhile, the high-stiffness point at the far end of the power assembly mounting layer refers to the reinforcing bolt hole or the edge of the adapter plate furthest from the fuselage in the area connecting the motor nacelle and the wing body. This location exhibits extremely high local stiffness and concentrated mass, allowing the force pulse to generate a very short contact time, thereby obtaining high-frequency energy across a wide frequency band. This effectively excites higher-order local modes and bending-torsional coupling modes constrained by local structures.

[0077] For each target excitation location, excitation is applied to each target excitation location using any of the following devices: a force hammer with a force sensor, a programmable automatic drop hammer system, a spring excitation gun, a non-contact acoustic excitation array, or a non-contact electromagnetic pulse exciter.

[0078] In practice, one or more of the above locations can be selected for multi-reference point excitation according to the focus of the test, and it is not limited to the limitations of a single device: a force hammer with an integrated force sensor can be used to obtain a precise force spectrum reference signal, a programmable automatic drop hammer can be used to ensure the repeatability of the striking force, or a spring excitation gun can be used to implement pulse excitation in a narrow gap, or even a non-contact acoustic excitation array or electromagnetic pulse exciter can be used to completely eliminate the influence of additional mass when conditions are limited.

[0079] The applied excitation direction is perpendicular to the surface of the wing under test. The core purpose of the hammer strike is to inject the widest possible bandwidth and highest energy excitation force into the structure in a single transient pulse. Perpendicular striking ensures the most direct and efficient normal momentum transfer between the hammer and the wing surface, allowing the force sensor to fully capture a pure force pulse signal without tangential interference. If an oblique strike is used, not only will frictional noise and force signal distortion be introduced due to tangential sliding between the hammer and the skin, but the tilted impact may also cause the hammer's rebound trajectory to deviate from the normal direction, disrupting the zero-mass boundary condition of "instantaneous detachment".

[0080] For each excitation process, the excitation signal during the current excitation process and the response signal at each target measurement point location in each feature segment are collected, including: If a force hammer with a force sensor is used to apply the excitation, then for each excitation process, the rising edge of the pulse signal of the force sensor is used to trigger the acquisition of the excitation signal in the current excitation process and the response signal of each target measurement point in each feature segment.

[0081] When applying modal excitation to an eVTOL wing using a force hammer equipped with a force sensor, the target excitation location is not randomly selected or limited to a single impact point. For a single wing, in actual testing, operators typically follow a predetermined multi-reference point impact testing strategy, sequentially striking two to three key locations that have undergone dynamic evaluation, such as the trailing edge of the wing and the high-stiffness point at the far end of the power assembly mounting layer.

[0082] For example, the wingtip trailing edge is struck several times with appropriate force to allow the broadband pulse energy generated there to propagate fully along the spanwise and chordwise directions of the wing, thereby effectively exciting the low-order overall bending and large-range torsional modes of the wing. Subsequently, the testers move to the mounting bracket or the edge of the stiffener with the greatest structural stiffness in the motor mounting area and strike it again to concentrate energy to excite the high-order local vibration modes that are constrained in a limited area due to local mass concentration and stiffness step.

[0083] Each excitation location is excited sequentially and independently in a single test. Finally, the multi-reference point modal parameter identification algorithm running on the computing control terminal mathematically fuses and solves multiple sets of frequency response function data from different excitation locations. This avoids the possibility of single-point excitation falling into the blind zone of high-order mode nodal lines, while obtaining a complete modal set covering the global and local features of the wing.

[0084] For each specific impact operation, the hammer strike direction is kept perpendicular to the surface of the wing under test. Perpendicular impact ensures that the force sensor fully captures a pure force pulse signal without tangential interference, and the hammer head instantly bounces off the normal under the structural reaction force, maintaining a zero-added-mass boundary. Synchronous excitation of bending and torsional modes is achieved by selecting excitation locations far from the elastic axis, such as the trailing edge of the wing, so that the vertical impact force naturally forms an effective lever arm relative to the torsional center of the structure. Upon hammer impact, the crystal inside the built-in force sensor is instantaneously compressed, generating a charge signal. Its voltage rises sharply within microseconds, forming a steep pulse rising edge.

[0085] The data acquisition system uses the moment when the rising edge crosses the preset voltage threshold as the hardware trigger reference. Without software intervention, it synchronously sends latching commands to all acquisition channels to initiate parallel analog-to-digital conversion and recording of the time-domain waveform of the excitation force signal output by the force sensor itself, as well as the time-domain waveform of the response signal of the acceleration sensor at all target measurement points distributed in each characteristic layer.

[0086] Once the striking action is completed, the hammerhead detaches instantly from the skin surface under the elastic reaction force of the wing structure, without remaining on the wing in any mechanical connection to continue participating in the subsequent free decay vibration process. This physical detachment characteristic ensures that the boundary conditions and mass distribution of the wing completely return to their true free state without interference from the testing equipment throughout the entire response phase after the hammerhead leaves, fundamentally eliminating the additional mass and stiffness errors introduced by traditional exciters and their push rod connections.

[0087] The above process is illustrated using a test example of the wing of a scaled-down eVTOL demonstrator aircraft: After completing the sensor topology setup, the tester, holding a modal force hammer with a built-in high-sensitivity force sensor, first stood at the first target excitation position approximately 0.1 meters from the wingtip on the trailing edge of the wing. The hammer handle angle was adjusted so that the hammer's trajectory formed a 90-degree angle with the upper surface of the wing, and five consecutive strikes were performed with moderate force, with approximately 30 seconds between each strike to allow for sufficient decay of the previous response. At each strike, the rising edge of the force sensor signal at the moment of contact triggered the data acquisition front-end. All acceleration sensors located in the root reinforcement layer, smooth mid-section transition layer, power component mounting layer, and wingtip free end layer synchronously started acquiring data on the same clock edge, recording the global vibration decay time history induced by the strike perpendicular to the wing surface at the trailing edge. After the first set of taps, the testers changed their position to the high-stiffness point at the far end of the aluminum alloy adapter bracket on the power assembly mounting layer on the lower surface of the wing. They adjusted the tapping direction to a 90-degree angle with the skin and tapped five more times. The same trigger mechanism then initiated the second round of synchronous data acquisition, capturing the high-frequency response signal dominated by the local vibration mode at the motor mounting location. After all tapping operations were completed, the time-domain data from all acquisition channels were uploaded to the computing control terminal. Because the hammerhead instantly bounced off the wing after each tap, both sets of data were obtained without additional mechanical constraints. This ensured that the input-output relationship used for subsequent frequency response function calculations and modal parameter identification strictly followed the true free-state dynamic characteristics of the wing structure. Consequently, the reconstructed first few bending-torsional coupling mode shapes exhibited extremely high physical fidelity in terms of frequency, damping ratio, and deformation spatial distribution.

[0088] Regarding step S16, based on the location of each target measurement point, the excitation signal of each excitation process, and the response signal, the target mode shape of the wing under test is constructed. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion, and their coupled deformation.

[0089] Step S16 serves as the data processing and mode visualization culmination of the entire testing process, transforming the time-domain response information scattered across various measurement points into a three-dimensional dynamic image reflecting the spatial vibration pattern of the wing.

[0090] First, the force input signals and acceleration response signals at each measuring point acquired during each excitation process are preprocessed using corresponding window functions, and then converted from time-domain waveforms to frequency-domain waveforms using Fast Fourier Transform. This allows for the calculation of the frequency response function of each target measuring point relative to different excitation reference points, forming a frequency response function matrix reflecting the global transmission characteristics of the wing. Based on this, a multi-reference-point modal parameter identification algorithm is used to perform pole search and residue estimation on this matrix within a preset target test frequency range, extracting the natural frequencies, damping ratios, and corresponding complex mode shape vectors covering the highest target mode order. Finally, the residue information in each mode shape vector is mapped back to the three-dimensional spatial coordinates of each target measuring point recorded in the previous steps, according to the acquisition channel number. This is then used to drive the finite element mesh nodes or geometric point clouds to generate deformation displacements consistent with the corresponding mode order through modal superposition or direct assignment, thereby reconstructing a target mode shape animation on the computing terminal that includes bending and torsional coupling components and accurately presents the high-frequency deformation characteristics of the local power component mounting area.

[0091] Specifically, based on the locations of each target measurement point, the excitation signals of each excitation process, and the response signals, the target mode shape of the wing under test is constructed, including: Step S161: Apply a rectangular window to the excitation signal in the time domain during each excitation process, and apply an exponentially decaying window to the response signal in the time domain during each excitation process. Step S162: Perform fast Fourier transform on the windowed target excitation signal and target response signal in each excitation process to obtain the frequency response function matrix of each target measurement point position relative to the target excitation position; Step S163: The multi-reference point modal parameter identification algorithm is called to process the frequency response function matrix and extract the poles and residues of the highest target modal order of the wing under test within the target test frequency range; the target test frequency range is determined based on the preset development requirements of the aircraft on which the wing under test is located and the minimum test frequency corresponding to each flutter analysis frequency band. Step S164: Based on the positions of each target measurement point and the poles and residues of the highest target modal order, construct the target mode shape of the wing under test. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation.

[0092] The phrase "including at least one of bending, torsion, and their coupled deformations" encompasses all possible mode shapes from low-order to high-order modes. Low-order modes typically manifest as a single deformation pattern dominated by overall wing bending or spanwise torsion. As the order increases, bending and torsional components gradually mix spatially, forming a complex deformation pattern with strong coupling between the two. Furthermore, some local modes may exhibit deformation characteristics dominated by chordal bending or oscillation. Regardless of the specific order of the mode shape, it falls within the scope of this description. In terms of quantity, regarding "first-order or multiple-order," when the mode of interest is only first-order, a first-order mode shape is output; when the mode of interest includes several orders, a corresponding number of multiple-order mode shapes are output.

[0093] In the digital signal processing stage, a rectangular window is applied to the time-domain excitation signal (such as a force signal). Its function is to precisely extract the main portion of the transient pulse energy generated at the moment of impact along the time axis, while simultaneously forcing the background noise range of the force sensor output before and after impact to zero. Because the hammer pulse width is extremely narrow and the signal energy is highly concentrated, the non-zero range of the rectangular window only needs to cover the brief period from the pulse rise to the amplitude decaying back to the baseline noise level. Data outside the window is completely zeroed, thus completely eliminating environmental electromagnetic interference and the contamination of the frequency domain force spectrum by the static bias of the acquisition circuit while ensuring the integrity of the pulse waveform. An exponentially decaying window is applied to the time-domain acceleration response signal. This window function exhibits a curve shape that smoothly decays from 1 to 0 in the time domain. Multiplying it by the response signal is equivalent to introducing known, artificially calculated damping during the free decaying vibration of the structure. Because the Fast Fourier Transform (FFT) implicitly assumes a periodic extension of a finite-length sampled sequence, if the response signal has not decayed to near zero by the end of a single sampling period, the extended signal will exhibit amplitude jumps at the period boundaries, leading to severe energy leakage, manifested as broadening of spectral peaks and uplift of sidelobes. The introduction of an exponential decay window forces the wing's free-decaying vibration signal to be compressed to near zero before the end of a single sampling period, eliminating the discontinuity effect at the beginning and end of the periodic extension, thus achieving a sharp and symmetrical resonance peak shape at the spectral level.

[0094] The computational control terminal performs a Fast Fourier Transform (FFT) on the excitation force target signals and acceleration target response signals at each measuring point after applying the aforementioned dual-window processing. This maps the discrete-time sequence to the frequency domain, obtaining the force autospectrum, response autospectrum, and the cross-spectral density function between them. Based on this, frequency response function estimation methods such as H1 estimation or Hv estimation are used to calculate the frequency response function of each target measuring point relative to the current excitation reference point channel by channel. The frequency response functions of all measuring points and all reference points are then assembled into a complete frequency response function matrix, with rows corresponding to response measuring points and columns corresponding to excitation reference points. Each column in this matrix represents the frequency domain response characteristics of all measuring points under the same reference point excitation, while each row reflects the transmission characteristics of a fixed measuring point under different reference point excitations.

[0095] Subsequently, a multi-reference point modal parameter identification algorithm is invoked to perform system pole search and modal parameter fitting on the frequency response function matrix within the target test frequency range jointly determined by the aircraft's pre-defined development requirements and flutter analysis frequency band. Taking the PolyMAX algorithm as an example, it constructs a right-matrix fractional model of the frequency response function, establishes denominator and numerator polynomials in the frequency domain, and plots steady-state diagrams using the statistical stability of pole estimation at different model orders. Within the target frequency band, the algorithm automatically selects physical poles that remain stable as the model order increases in three dimensions: frequency, damping ratio, and participation factor. Each stable pole corresponds to the first-order true intrinsic mode of the wing. For each confirmed physical pole, the algorithm simultaneously extracts its corresponding residue matrix. The residue contains the complex amplitude ratio and phase relationship of each response measurement point relative to each reference point under that mode, essentially defining the full-wing spatial deformation mode of that mode. Since the previous spatial anti-aliasing formula has ensured that the density of physical measurement points is strictly higher than the shortest spatial wavelength corresponding to the highest target modal order, the PolyMAX algorithm extracts the residue matrix before... When dealing with poles and residues, there is no need to rely on model extension techniques that suffer from spatial interpolation distortion; instead, the quantity strictly satisfies... The parameters are fitted using the physical degrees of freedom response matrix composed of real accelerometers.

[0096] Finally, based on the three-dimensional spatial coordinates of each target measurement point in the global coordinate system, which were accurately recorded in the previous steps, the residue vector of each mode is bound to the spatial coordinates one by one according to the channel number. For any given mode, the residue amplitude at the measurement point represents the relative displacement scaling factor of that point in the mode shape, while the residue phase characterizes the lead or lag relationship between its motion direction and the reference excitation. The computing control terminal uses the mode shape information at these discrete measurement points as control points to drive the surface mesh or node cloud of the three-dimensional geometric model of the wing under test to generate corresponding proportional deformation displacements. Based on the topological connection relationship between each point, smooth rendering and dynamic frame interpolation are performed, and finally, a three-dimensional mode shape animation that can be freely rotated, scaled, and sectioned for observation is generated on the display. This animation not only fully presents the overall bending and torsional coupling deformation envelope of the wing across its entire span, but also, because its data source comes from a measured physical lattice that satisfies the Nyquist space sampling density, it can clearly reproduce the dense spatial fluctuation details unique to high-frequency modes in local stiffness step regions such as the power component mounting layer and the stiffener termination line, thus drawing high-frequency local coupling deformation attitudes without any spatial distortion.

[0097] Taking the data processing of a certain type of eVTOL wing during the extraction of the first six modes as an example: After calculation using the anti-aliasing formula, 19 accelerometers were deployed on the power component mounting layer and 16 sensors were deployed on the wingtip free end layer. A force hammer was used to sequentially strike the trailing edge of the wingtip and the high-stiffness point at the far end of the motor support, with five sets of valid data collected from each of the two excitations. First, each force signal was truncated by a rectangular window of 0.1 seconds, retaining only the main pulse and its adjacent zero region. Each response signal was multiplied by an exponential window with an adjusted attenuation coefficient, so that the free vibration, which originally required 8 seconds to decay, had decreased to less than one percent of its peak value by the end of the 2-second sampling window. After performing a fast Fourier transform on the windowed data, the system generated a total of 35 frequency response function curves at two reference points, forming a 35-row by 2-column frequency response function matrix. The PolyMAX algorithm analyzes the matrix within the target frequency band of 0 to 80 Hz. The steady-state plot shows a clear sequence of stable poles at 12.1 Hz, 22.3 Hz, 31.8 Hz, 42.5 Hz, 51.2 Hz, and 58.7 Hz, with damping ratios identified as 2.1%, 1.8%, 3.2%, 2.5%, 4.1%, and 3.6%, respectively. After extracting the residue vectors of each order, they are mapped to the known three-dimensional coordinates of 35 measurement points, driving the geometric model to generate dynamic deformation. The final sixth-order bending-torsional coupling mode animation shows that near the trailing edge of the motor in the power component mounting layer, the wing skin exhibits three local bulges and depressions arranged along the span, forming a dense wave-like deformation with a spatial wavelength of approximately 0.12 meters. The average spacing between the measurement points in this area is 0.09 meters, which meets the Nyquist requirement of at least two sampling points per wavelength. Therefore, this local high-frequency deformation feature is completely captured and presented on the screen without any omissions or phase ambiguity.

[0098] In summary, the embodiments of this application extract the mass concentration block and stiffness step block of the wing under test along the spanwise direction and divide it into characteristic segments with independent dynamic properties. Combined with the aircraft development requirements and flutter analysis frequency band, the highest target mode order is determined. Then, based on the Nyquist space sampling law, the minimum number of measurement points is quantitatively configured for each segment. The excitation signal and the response signal of each measurement point are obtained by sequential excitation and synchronous acquisition. Finally, the target mode shape is reconstructed. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation. This application's embodiments transform the measurement point layout from an experience-based equidistant arrangement to a quantitative decision driven by both structural physical characteristics and sampling theorems. This fundamentally eliminates sampling aliasing and modal loss caused by the shortening of the spatial wavelength of higher-order vibration modes. At the same time, the excitation method of instantaneous detachment after excitation eliminates the interference of the added mass of the traditional exciter on the natural frequency of the lightweight wing. This ensures that the identified modal parameters truly reflect the dynamic characteristics of the wing in its free state, significantly improving the completeness and data confidence of the identification of complex bending-torsional coupling vibration modes and local high-frequency modes. This provides high-fidelity experimental data support for flutter boundary prediction and structural safety assessment.

[0099] The embodiments of this application refine the identification of mass concentration blocks and stiffness step blocks by using the correlation of physical properties of finite element meshes and the statistics of extreme points and jump points of spanwise mass density gradient function and bending stiffness gradient function. This provides a clear calculation basis for the delineation of feature regions and the determination of the number of feature segments, eliminating the subjectivity and inconsistency of manual experience-based division.

[0100] The embodiments of this application clarify the process of determining the highest target modal order by covering multiple source frequency requirements such as flight control bandwidth, vibration and noise control, and airworthiness certification, and statistically counting the number of natural frequency characteristic values ​​in the finite element model accordingly, so as to ensure that the test range fully meets all top-level indicators and does not omit key modes.

[0101] This application embodiment concretizes the application of the Nyquist spatial sampling theorem by assigning differentiated spatial oversampling ratios based on the structural distribution characteristics of each layer, and calculating the minimum number of measurement points by combining the anti-leakage redundancy constant with the rounding up formula, so that the degree of measurement point densification is precisely matched with the local structural complexity. Even when facing multiple power mounting layers or complex reinforced structures of the wing, it can still adaptively generate a point layout scheme that ensures no aliasing sampling.

[0102] This application proposes constraints on the spatial topological distribution of measurement points within the layer, such as avoiding the flexible center of the skin, maximizing the spanwise and chordal envelopes, and high-density layout of the dynamic mounting layer. It is also compatible with both non-contact optical measurement and contact accelerometer response acquisition methods. This not only filters out the adverse effects of local skin bulging noise on the torsional vibration mode signal from a physical perspective, but also enables the hardware selection of the test system to flexibly adapt to different test conditions.

[0103] This application embodiment clarifies the selection strategy for multi-reference point excitation positions and the triggering and acquisition methods of replaceable excitation devices such as force hammers, automatic falling hammers, and electromagnetic pulse excitation. By sequentially striking the trailing edge of the wing and the high-stiffness point at the far end and triggering synchronous acquisition with hardware on the rising edge of the pulse, it avoids insufficient modal excitation caused by single-point excitation falling into the nodal line, while ensuring nanosecond-level synchronization accuracy between each excitation and the multi-channel response.

[0104] In this embodiment, the signal processing flow before mode shape reconstruction is refined to apply a rectangular window to the excitation signal to extract pure pulses, apply an exponentially decaying window to the response signal to eliminate periodic extension leakage, construct the frequency response function matrix through fast Fourier transform, and extract the pre-modal parameter using a multi-reference point modal parameter identification algorithm. The order poles and residues enable the final generated target mode shape animation to achieve a high fidelity level suitable for airworthiness certification in terms of frequency identification accuracy, damping ratio estimation stability, and spatial deformation detail representation.

[0105] Since the electronic device described in this embodiment is the electronic device used to implement the information processing method in the embodiments of this application, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the information processing method described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here.

[0106] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0111] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for testing the modal characteristics of an aircraft wing, characterized in that, The method includes: Based on the three-dimensional structural model of the wing under test, the mass concentration block and stiffness step block in the wing under test are determined. Based on the distribution positions of the mass concentration blocks and the stiffness step blocks on the wing under test, k characteristic segments arranged along the spanwise direction of the wing under test are determined; k is a positive integer. Based on the pre-set development requirements and flutter analysis frequency band of the aircraft on which the wing under test is located, the highest target mode order that needs to be extracted from the wing under test is determined; Based on the highest target modal order and the structural distribution characteristics of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined according to the Nyquist space sampling theorem. Excitation is applied to each target excitation position on the wing under test one by one. For each excitation process, the excitation signal in the current excitation process and the response signal of each target measurement point position in each feature segment are collected. The number of target measurement point positions in each feature segment is greater than or equal to the minimum number of measurement points corresponding to the feature segment, and satisfies the preset conditions of the intra-layer spatial topology distribution corresponding to the feature segment. Based on the locations of each target measurement point, the excitation signals and response signals of each excitation process, the target mode shape of the wing under test is constructed. The target mode shape is a first-order or multi-order mode shape that includes at least one of bending, torsion and their coupled deformation.

2. The aircraft wing modal testing method as described in claim 1, characterized in that, Based on the three-dimensional structural model of the wing under test, the mass concentration block and stiffness step block in the wing under test are determined, including: The three-dimensional structural model of the wing under test is preprocessed using the finite element method to generate an initial element mesh. Associating the initial cell mesh with its corresponding physical properties yields the target cell mesh; Based on the target cell mesh, determine the linear mass density distribution curve and bending stiffness distribution curve along the span of the wing under test; The mass concentration blocks in the wing under test are determined based on the linear mass density distribution curve. The stiffness step block in the wing under test is determined based on the bending stiffness distribution curve.

3. The aircraft wing modal testing method as described in claim 1, characterized in that, Based on the distribution positions of the mass concentration blocks and the stiffness step blocks on the wing under test, k characteristic segments arranged along the spanwise direction of the wing under test are determined, including: Based on the distribution positions of the mass concentration blocks and the stiffness step blocks on the wing under test, the mass density gradient function and bending stiffness gradient function of the wing under test along the span are determined. The sum of the number of extreme points in the mass density gradient function and the number of jump points in the bending stiffness gradient function is determined as the number of characteristic segments arranged along the span of the wing under test.

4. The aircraft wing modal testing method as described in claim 1, characterized in that, The pre-set development requirements include at least flight control bandwidth requirements, vibration and noise control requirements, and airworthiness certification requirements. Based on the pre-defined development requirements and flutter analysis frequency band of the aircraft containing the wing under test, the highest target mode order to be extracted from the wing under test is determined, including: Based on the preset development requirements of the aircraft containing the wing under test and the minimum test frequency corresponding to each flutter analysis frequency band, the target test frequency range of the aircraft is determined; the target test frequency range covers the minimum test frequency corresponding to each preset development requirements and flutter analysis frequency band. Finite element analysis is performed on the target test frequency range to determine the number of natural frequency eigenvalues ​​of the target test frequency range, and the number of natural frequency eigenvalues ​​is determined as the highest target modal order.

5. The aircraft wing modal testing method as described in claim 1, characterized in that, Based on the highest target modal order and the structural distribution characteristics of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined according to the Nyquist space sampling theorem, including: Based on the structural distribution characteristics of each feature segment and the preset correlation, the spatial oversampling ratio corresponding to each feature segment is determined; the preset correlation characterizes the correspondence between the structural distribution characteristics and the spatial oversampling ratio; the preset correlation is determined based on the Nyquist spatial sampling theorem. Based on the highest target modal order, the preset leakage redundancy constant, and the spatial oversampling ratio of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined.

6. The aircraft wing modal testing method as described in claim 5, characterized in that, Based on the highest target modal order, the preset leakage redundancy constant, and the spatial oversampling ratio of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined, including: Based on the preset formula, the highest target modal order, the preset leakage prevention redundancy constant, and the spatial oversampling ratio of each feature segment, the minimum number of measurement points corresponding to each feature segment is determined. The preset formula is: in, Indicates the first The minimum number of measurement points corresponding to each feature layer segment Indicates to Round up. Indicates the first Spatial oversampling ratio corresponding to each feature segment This indicates the highest target modal order. This represents the preset leakage prevention redundancy constant.

7. The aircraft wing modal testing method as described in claim 1, characterized in that, The preset conditions for the spatial topology distribution within the layer include: avoiding the target measurement point location being in the flexible central region of the wing skin; maximizing the envelope of adjacent target measurement point locations in the spanwise and chordwise directions of the wing under test; and ensuring that the density of target measurement point locations in the feature layer segment containing the power mount is greater than the density of target measurement point locations in the feature layer segment not containing the power mount.

8. The aircraft wing modal testing method as described in claim 7, characterized in that, For each excitation process, the response signals of each target measurement point in each feature segment during the current excitation process are collected, including: For each excitation process, a non-contact three-dimensional laser Doppler vibrometer or a high-speed binocular stereo vision camera is used to collect the response signals of each target measurement point in each feature segment during the current excitation process; or, Before applying excitation to each target excitation location on the wing under test one by one, the method further includes: According to the preset conditions of the intra-layer spatial topology distribution corresponding to each feature segment, the target measurement point position of the accelerometer is determined in each feature segment, and the accelerometer is rigidly attached at each target measurement point position by aviation beeswax or glue. For each excitation process, the response signals of each target measurement point in each feature segment during the current excitation process are collected, including: for each excitation process, the response signals of each target measurement point in each feature segment during the current excitation process are collected by the accelerometer at each target measurement point.

9. The method for modal testing of an aircraft wing as described in claim 1, characterized in that, The target excitation location includes at least one of the following locations: the trailing edge of the wing and the high-stiffness point at the far end of the power assembly mounting layer. For each target excitation location, excitation is applied to each target excitation location using any of the following devices: a force hammer with a force sensor, a programmable automatic drop hammer system, a spring excitation gun, a non-contact acoustic excitation array, or a non-contact electromagnetic pulse exciter. The applied excitation direction is perpendicular to the surface of the wing under test; For each excitation process, the excitation signal during the current excitation process and the response signal at each target measurement point location in each feature segment are collected, including: If a force hammer with a force sensor is used to apply the excitation, then for each excitation process, the rising edge of the pulse signal of the force sensor is used to trigger the acquisition of the excitation signal in the current excitation process and the response signal of each target measurement point in each feature segment.

10. The method for modal testing of an aircraft wing as described in claim 1, characterized in that, Based on the locations of each target measurement point, the excitation signals and response signals of each excitation process, the target mode shape of the wing under test is constructed, including: A rectangular window is applied to the time-domain excitation signal in each excitation process, and an exponentially decaying window is applied to the time-domain response signal in each excitation process. Perform a fast Fourier transform on the windowed target excitation signal and the target response signal in each excitation process to obtain the frequency response function matrix of each target measurement point position relative to the target excitation position; The frequency response function matrix is ​​processed by a multi-reference point modal parameter identification algorithm to extract the poles and residues of the highest target modal order of the wing under test within the target test frequency range; the target test frequency range is determined based on the preset development requirements of the aircraft on which the wing under test is located and the minimum test frequency corresponding to each flutter analysis frequency band. Based on the locations of each target measurement point and the poles and residues of the highest target mode order, the target mode shape of the wing under test is constructed.

Images