A modal inversion-based aero-engine rotor static stiffness testing method

CN122671103APending Publication Date: 2026-09-01NORTHEASTERN UNIV CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610803309.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-06-05
Publication Date
2026-09-01

AI Technical Summary

Technical Problem

[0003]有鉴于此,本发明的目的在于提出一种基于模态反演的航空发动机转子静态刚度测试方法,以克服传统静刚度测试方法成本高、实现困难的缺陷,为航空发动机转子提供一种低成本、易实施且精度满足工程需求的刚度测试手段

Benefits of technology

本发明以低成本的模态试验替代传统静态加载,结合静态刚度反演的参数化转子动力学模型和遗传优化算法,实现了航空发动机转子刚度的准确反演识别。在模拟转子中,模态反演刚度与准静态试验结果最大相对误差仅为-3.14%;在高压转子中,各测点反演刚度与仿真结果误差均小于4.5%。该方法成本低、易实施,对复杂结构适应性强,为转子静态刚度测试提供了新的有效途径,具有显著的工程实用价值。

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122671103A_ABST
    Figure CN122671103A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of aero-engine rotor dynamics and stiffness testing technology, and discloses a method for testing the static stiffness of aero-engine rotors based on modal inversion. It replaces traditional static loading with low-cost modal testing, and combines a parameterized rotor dynamics model and genetic optimization algorithm for static stiffness inversion to achieve accurate inversion identification of aero-engine rotor stiffness. In simulated rotors, the maximum relative error between the modal inversion stiffness and the quasi-static test results is only -3.14%; in high-pressure rotors, the error between the inverted stiffness at each measuring point and the simulation results is less than 4.5%. This method is low-cost, easy to implement, and highly adaptable to complex structures, providing a new and effective approach for rotor static stiffness testing, and has significant engineering practical value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of aero-engine rotor dynamics and stiffness testing technology, and in particular to a method for testing the static stiffness of aero-engine rotors based on modal inversion. Background Technology

[0002] The stiffness distribution of an aero-engine rotor directly affects its dynamic response characteristics and the accuracy of tip clearance control between the rotor and stator. Insufficient or uneven stiffness distribution can lead to excessive rotor deformation and intensified vibration, and in severe cases, even rubbing failure, threatening the structural integrity and operational safety of the engine. Traditional rotor static stiffness testing is mostly based on the force-deformation relationship under static loading. A known static load is applied to the rotor through a hydraulic actuator or weights, and the deformation is obtained using displacement sensors, thereby calculating the equivalent bending stiffness or support stiffness. This method is simple in principle, but it has significant limitations when dealing with the increasingly complex aero-engine rotors: on the one hand, it requires the design and construction of specialized loading fixtures, resulting in high testing costs and long cycles; on the other hand, the high-pressure rotors and other components of real aero-engines have compact structures and complex assembly states, making it difficult to install static loading devices and restricting the arrangement of measuring points, making it difficult to accurately implement static stiffness testing based on the force-deformation principle, and even impossible to perform under certain operating conditions. Therefore, how to obtain the static stiffness characteristics of aero-engine rotors in a low-cost and highly feasible manner has become an urgent problem to be solved in the field of rotor dynamics. Summary of the Invention

[0003] In view of this, the purpose of this invention is to propose a static stiffness testing method for aero-engine rotors based on modal inversion, so as to overcome the shortcomings of traditional static stiffness testing methods, which are costly and difficult to implement, and provide a low-cost, easy-to-implement stiffness testing method for aero-engine rotors with accuracy that meets engineering requirements.

[0004] The technical solution adopted in this invention is as follows: A method for testing the static stiffness of an aero-engine rotor based on modal inversion, comprising the following steps: Step 1: Obtain the quasi-static stiffness reference and modal parameters of the rotor: Perform quasi-static stiffness test and modal test on the simulated rotor respectively, and obtain the bending stiffness value of the simulated rotor under graded static load as a reference. At the same time, measure the first three bending natural frequencies of the rotor and the corresponding mode shapes. Step 2: Construct a parametric rotor dynamics model for static stiffness inversion. Discretize the rotor along the axial direction into a finite element system containing beam elements, concentrated mass points, and spring-damped supports. Based on the distribution characteristics of the shaft segment, disc, flange, connection interface, and support position in the actual rotor structure, divide the rotor into multiple parameter identification segments corresponding to the sensitivity to stiffness changes, and set equivalent stiffness parameters that can be updated independently for each parameter identification segment. Based on the arrangement of measurement points in the modal test, establish a corresponding mapping relationship between the test measurement points and the nodal degrees of freedom of the parametric rotor dynamics model, so that the calculated bending natural frequencies and mode shapes can be compared with the test modal parameters in a unified coordinate and unified degree of freedom space. Step 3, Genetic Optimization Inversion: Using the bending natural frequency and mode shape obtained in Step 1 as the joint inversion target, the stiffness parameters of each parameter identification section of the parameterized rotor dynamics model in Step 2 are subjected to global iterative optimization with physical constraints. During the optimization process, an objective function is constructed based on the frequency error, mode shape deviation, and coordination relationship of the changes in section parameters between the calculated modal parameters and the experimental modal parameters. Combined with the total rotor mass constraint and the range constraint of the section parameter values, the stiffness distribution of the parameter identification section is updated generation by generation until the objective function meets the preset convergence condition, and the rotor modal inversion stiffness consistent with the measured modal characteristics is obtained.

[0005] Step 1 specifically includes the following steps: Step 11: Apply graded static loads radially at different axial positions of the simulated rotor and extract the equivalent bending stiffness values ​​at multiple cross sections of the simulated rotor. Step 12: Perform modal tests on the same simulated rotor, obtain the frequency response function at each measuring point, and identify the first three bending natural frequencies and mode shapes.

[0006] In step 11, a graded static load is applied radially at different axial positions of the simulated rotor, with a load range of 0 N to 5000 N and a step size of 500 N. The radial deformation at each measurement point is recorded synchronously using a dial indicator to obtain a force-deformation curve. The equivalent bending stiffness value of each section is extracted based on the slope of the force-deformation curve and used as the quasi-static stiffness reference data. After completing the initial azimuth angle test, the simulated rotor is rotated sequentially by 90°, 180°, and 270°, and the loading and measurement are repeated. The average value of the multiple measurement results is taken as the equivalent bending stiffness of the section.

[0007] In step 12, multiple effective hammer blows are performed at each excitation point, and the average frequency response function is taken. The frequency response function data of each measurement point are processed by the modal analysis system, and the first three bending natural frequencies and mode shapes are identified by the least squares complex frequency domain method.

[0008] Step 2 specifically includes the following steps: The rotor shaft segment was discretized using Timoshenko beam elements, simplifying the shaft segment into beam elements, the disc and flange into concentrated mass points, and the bearing support into linear spring-damped elements. Based on the rotor geometry, abrupt changes in cross-section, disc-shaft connection, flange connection, and support positions, the rotor was divided into several parameter identification segments along the axial direction. Each parameter identification segment was assigned independent cross-sectional parameters and / or equivalent material parameters to form a segmented parameterized model that could characterize the non-uniform distribution of local stiffness in the rotor. Based on the sensor placement and excitation positions in the modal test, a spatial correspondence between the test measurement points and the nodes of the finite element model was established. The modal shapes obtained from the test were mapped to the corresponding node degrees of freedom of the finite element model after interpolation, normalization, and degree-of-freedom transformation to ensure that the test modal parameters and the calculated modal parameters were calculated for error on the same evaluation basis. Rotor discretization N Each node constructs a mass matrix. M Stiffness matrix K and damping matrix C The dynamic equations of the rotor system: in, F For external force vectors, q is the nodal displacement vector.

[0009] Step 3 specifically includes: Step 31, Variable Definition: Define the cross-sectional area of ​​each sensitive shaft segment of the rotor. A j The elastic modulus is set as the optimization variable, and the optimization variable is used to characterize the local bending stiffness distribution of the rotor in different sections along the axial direction. The optimization variables of each parameter identification section are set with independent upper and lower limits according to the corresponding structural position to reflect the differences in actual structural size and material properties of different sections. Step 32, Objective Function Construction: Define and compute the objective function. F ( A The weighted sum of squares of the relative errors between the simulated natural frequencies and the experimental natural frequencies is: in, δ i For the first i The weighting coefficients for the first-order frequencies are all set to 1; simultaneously, the total rotor mass and cross-sectional area are used as constraints, where... ρ For material density, L j For the first j The beam segment length is determined to ensure physical consistency of the model. The rotor shaft is discretized into... N Each beam segment, fsim,i ( A () is the first eigenvalue obtained by solving the eigenvalues ​​based on the beam model. i First bending frequency, f exp,i The first one measured in the experiment i First bending frequency; Step 33, Optimization: A genetic algorithm is used to perform a global search on the optimization variables. The parameter vector is updated generation by generation through selection, crossover, and mutation operations. In each iteration, each candidate parameter vector is substituted into the parameterized dynamic model established in Step 2 to solve for eigenvalues, obtain the corresponding bending natural frequencies and mode shapes, and calculate the matching degree of the candidate parameter vectors with the test modal parameters based on the objective function. Candidate parameter vectors that meet the constraints and have a better objective function are retained and evolved until the objective function reaches the preset minimum tolerance value, or the rate of change of the optimal objective function is less than a set threshold for several consecutive generations. Then, the iteration stops and the current optimal parameter vector is output as the identification result of the rotor stiffness distribution parameters, and the modal inversion stiffness is obtained accordingly.

[0010] The specific process of step 33 includes: Step 331: Initialize the population: Within the preset physical feasible domain of the cross sections and elastic moduli of each shaft segment of the rotor system, randomly generate P initial parameter vector populations that satisfy the above mass constraints. Step 332, Fitness Assessment: Substitute the parameter vector of each individual in the population into the parameterized dynamics model to solve for the eigenvalues ​​and obtain the first... i 1st-order simulated bending natural frequency f sim,i And calculate the corresponding objective function value. F ( A ); 1 / (1 + F ( A The fitness value is defined as the individual's fitness value, used to quantify how closely the individual approximates the experimental data. Step 333, Genetic Operator Operations: Based on fitness values, perform selection operations on individuals in the population, retaining individuals with high fitness; perform crossover and mutation operations on the selected samples to generate a new generation of parameter vector population, and perform quality constraint verification on the new generation population, removing individuals that do not meet the constraints; Step 334, Iterative Convergence Determination: Repeat steps 332 to 333 until the... k Objective function of generation population F ( A When the preset minimum value is reached, or when the change in the objective function value for several consecutive generations is less than the set relative error threshold, the iteration stops and the current optimal parameter vector is output as the identification result of the stiffness distribution parameter.

[0011] The modal inversion stiffness calculation process includes: Based on the bending natural frequencies and mode shapes identified in step 3, when the excitation frequency approaches zero, the displacement frequency response function degenerates into a static displacement response, thereby establishing the rotor's modal displacement compliance matrix: in, For the first i Normalized mode shape of the first order; ω i For the first i The undamped natural frequency of each order is determined; the modal parameters obtained from each order identification are substituted into the modal displacement compliance matrix to obtain the equivalent static compliance matrix corresponding to the position of each measuring point or target section; and by inverting the equivalent static compliance matrix, the mapping from dynamic modal characteristics to static stiffness values ​​is realized, thereby obtaining the modal inversion stiffness of the rotor.

[0012] Compared with the prior art, the present invention has the following beneficial effects: This invention replaces traditional static loading with low-cost modal testing, combining a parameterized rotor dynamics model for static stiffness inversion with a genetic optimization algorithm to achieve accurate inversion and identification of aero-engine rotor stiffness. In simulated rotors, the maximum relative error between modal inversion stiffness and quasi-static test results is only -3.14%; in high-pressure rotors, the errors between inverted stiffness at each measurement point and simulation results are all less than 4.5%. This method is low-cost, easy to implement, and highly adaptable to complex structures, providing a new and effective approach for rotor static stiffness testing, and has significant engineering practical value. Attached Figure Description

[0013] Figure 1 This is a flowchart of a method for testing the rotor stiffness of an aero-engine based on modal inversion according to the present invention. Figure 2 A schematic diagram showing the structure of the simulated rotor specimen and the distribution of measurement points; Figure 3 The displacement-load curve test results are shown to simulate the rotor under different azimuth angles; (a) corresponds to measurement point 5, (b) corresponds to measurement point 6; Figure 4 The measured deflection curves of the rotor at various points under applied load are shown in the following figures: (a) Loading at measurement point 2; (b) Loading at measurement point 3; (c) Loading at measurement point 4; (d) Loading at measurement point 5; (e) Loading at measurement point 6; (f) Loading at measurement point 7. Figure 5 The figures show the comparison results of natural frequencies and mode shapes between simulation and experiment; (a) is the first order; (b) is the second order; (c) is the third order. Figure 6 This is a schematic diagram of a simulated rotor dynamics model; Figure 7 (a) is a schematic diagram of the genetic optimization iterative process; (b) is a schematic diagram of the gene optimization algorithm; Figure 8 (a) is a finite element simulation model of the rotor of an aero-engine; (b) is a decomposed model of the main components; (c) is a finite element model of the whole; and (d) is a partial structural diagram of the double lug stop and bolts. Figure 9 The following is a comparison of the two methods for simulating the rotor and the simulated deflection curve; (a) is the measurement point 5, and (b) is the measurement point 6. Figure 10 A schematic diagram of the high-pressure rotor beam element model and measurement point distribution for an aero-engine; Figure 11 The following is a comparison of the deformation of the high-voltage rotor with the simulation model results; (a) is measuring point 13; (b) is measuring point 26; (c) is measuring point 40.

[0014] Figure 12 This is a comparison diagram of the lateral stiffness distribution of high-pressure rotors. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0016] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] A method for testing the rotor stiffness of an aero-engine based on modal inversion includes the following steps: S1. Obtain the quasi-static stiffness reference and modal parameters of the rotor: Perform quasi-static stiffness test and modal test on the simulated rotor respectively, obtain the bending stiffness value of the simulated rotor under graded static load as a reference, and at the same time measure the first three bending natural frequencies of the rotor and the corresponding mode shapes. S2. Construct a parametric rotor dynamics model for static stiffness inversion. Discretize the rotor along the axial direction into a finite element system containing beam elements, concentrated mass points, and spring-damped supports. Based on the distribution characteristics of the shaft segment, disc, flange, connection interface, and support position in the actual rotor structure, divide the rotor into multiple parameter identification segments corresponding to the sensitivity to stiffness changes, and set equivalent stiffness parameters that can be updated independently for each parameter identification segment. Based on the arrangement of measurement points in the modal test, establish a corresponding mapping relationship between the test measurement points and the degrees of freedom of the nodes in the parametric dynamics model, so that the calculated bending natural frequencies and mode shapes can be compared with the test modal parameters in a unified coordinate and unified degree of freedom space. S3. Genetic Optimization Inversion: Using the bending natural frequency and mode shape obtained in step 1 as the joint inversion target, the stiffness parameters of each parameter identification section of the parameterized rotor dynamics model in step 2 are subjected to global iterative optimization with physical constraints. During the optimization process, an objective function is constructed based on the frequency error, mode shape deviation and coordination relationship of the changes in the section parameters between the calculated modal parameters and the experimental modal parameters. Combined with the total rotor mass constraint and the range constraint of the section parameter values, the stiffness distribution of the parameter identification section is updated generation by generation until the objective function meets the preset convergence condition, and the rotor modal inversion stiffness consistent with the measured modal characteristics is obtained.

[0018] S4. Stiffness Verification and Engineering Application: The modal inversion stiffness is compared and verified with the quasi-static stiffness benchmark of S1 and the ANSYS finite element simulation results to evaluate the identification accuracy. After the accuracy meets the requirements, the modal inversion stiffness method is extended to the stiffness evaluation of the high-pressure rotor of a real aero-engine.

[0019] S1 specifically includes the following steps: S11. Apply graded static loads radially at different axial positions of the simulated rotor, measure the radial deformation at the loading point using a displacement sensor, record the force-deformation curve, and extract the equivalent bending stiffness values ​​at multiple cross sections of the simulated rotor as quasi-static stiffness reference data. S12. Perform modal tests on the same simulated rotor, obtain the frequency response function at each measuring point, identify the natural frequencies and mode shapes of the first three bending modes, and record the frequency values ​​and mode shapes of each mode.

[0020] In step S11, a graded static load is applied radially at different axial positions of the simulated rotor, with a load range of 0 N to 5000 N and a step size of 500 N. The radial deformation at each measurement point is recorded synchronously using a dial indicator to obtain a force-deformation curve. The equivalent bending stiffness value of each section is extracted based on the slope of the curve and used as quasi-static stiffness reference data. To evaluate the circumferential consistency of the rotor, after completing the initial azimuth angle test, the rotor is rotated sequentially by 90°, 180°, and 270°, and the loading and measurement are repeated. The average value of the four measurement results is taken as the equivalent bending stiffness of the section.

[0021] In step S12, three effective hammer blows are performed at each excitation point, and the average frequency response function is taken to eliminate the influence of random noise. The frequency response function data of each measuring point are processed by a modal analysis system, and the first three bending natural frequencies and mode shapes are identified using the least squares complex frequency domain method. The identified first three bending natural frequencies are: first bending 118.75 Hz, second bending 215.77 Hz, and third bending 316.12 Hz. Each mode shape exhibits reasonable local variations at the support position and the mass disk, and the mode shape node positions are consistent with the rotor structural characteristics.

[0022] S2 specifically includes the following steps: The rotor shaft segment was discretized using Timoshenko beam elements, simplifying the shaft segment into beam elements, the disc and flange into concentrated mass points, and the bearing support into linear spring-damped elements. Based on the rotor geometry, abrupt changes in cross-section, disc-shaft connection, flange connection, and support positions, the rotor was divided into several parameter identification segments along the axial direction. Each parameter identification segment was assigned independent cross-sectional parameters and / or equivalent material parameters to form a segmented parameterized model that could characterize the non-uniform distribution of local stiffness in the rotor. Based on the sensor placement and excitation positions in the modal test, a spatial correspondence between the test measurement points and the nodes of the finite element model was established. The modal shapes obtained from the test were mapped to the corresponding node degrees of freedom of the finite element model after interpolation, normalization, and degree-of-freedom transformation to ensure that the test modal parameters and the calculated modal parameters were calculated for error on the same evaluation basis. Rotor discretization N Each node constructs a mass matrix. M Stiffness matrix K and damping matrix C The dynamic equations of the rotor system: in, F For external force vectors, q is the nodal displacement vector.

[0023] S3 specifically includes: Step 31, Variable Definition: Define the cross-sectional area of ​​each sensitive shaft segment of the rotor. A j The elastic modulus is set as the optimization variable, and the optimization variable is used to characterize the local bending stiffness distribution of the rotor in different sections along the axial direction. The optimization variables of each parameter identification section are set with independent upper and lower limits according to the corresponding structural position to reflect the differences in actual structural size and material properties of different sections. Step 32, Objective Function Construction: Define and compute the objective function. F ( A The weighted sum of squares of the relative errors between the simulated natural frequencies and the experimental natural frequencies is: in, δ i For the first i The weighting coefficients for the first-order frequencies are all set to 1; simultaneously, the total rotor mass and cross-sectional area are used as constraints, where... ρ For material density, L j For the first j The beam segment length is determined to ensure physical consistency of the model; the rotor shaft is discretized into... N Each beam segment, f sim,i ( A () is the first eigenvalue obtained by solving the eigenvalues ​​based on the beam model. i First bending frequency, f exp,i The first one measured in the experiment i First bending frequency; Step 33, Optimization: A genetic algorithm is used to perform a global search on the optimization variables. The parameter vector is updated generation by generation through selection, crossover, and mutation operations. In each iteration, each candidate parameter vector is substituted into the parameterized dynamic model established in Step 2 to solve for eigenvalues, obtain the corresponding bending natural frequencies and mode shapes, and calculate the matching degree of the candidate parameter vectors with the test modal parameters based on the objective function. Candidate parameter vectors that meet the constraints and have a better objective function are retained and evolved until the objective function reaches the preset minimum tolerance value, or the rate of change of the optimal objective function is less than a set threshold for several consecutive generations. Then, the iteration stops and the current optimal parameter vector is output as the identification result of the rotor stiffness distribution parameters, and the modal inversion stiffness is obtained accordingly.

[0024] Furthermore, the feasibility of the method of the present invention is verified through step S4; step S4 specifically includes the following steps: Step 41: Establish a finite element solid model consistent with the simulated rotor, perform static and modal analysis, and obtain the simulation stiffness and frequency; Step 42: Compare the quasi-static test stiffness, modal inversion stiffness, and simulation stiffness, calculate the error based on the quasi-static results, and verify the accuracy of the method.

[0025] Step 41 specifically includes: A three-dimensional solid finite element model consistent with the simulated rotor was established using finite element software. The model was meshed using solid elements. Contact elements were used to simulate the interface contact state at the mating surfaces of the two-stage disc-drum flange and the stop connection position. The bolt preload was applied using preload elements, and the bearing and coupling supports were simulated using spring-damping elements. Static and modal analyses were performed on the finite element model. In the static analysis, concentrated loads were applied sequentially at each measurement point, and the radial displacement of the corresponding nodes was calculated. The equivalent bending stiffness of each measurement point was obtained based on the force-deformation relationship. In the modal analysis, the natural frequencies and mode shapes of the rotor in the free-free state were calculated, and the first three bending frequencies and their corresponding mode shapes were obtained.

[0026] Specifically, one embodiment is as follows: A three-dimensional solid finite element model consistent with the simulated rotor was established using the finite element software ANSYS. The model was meshed using Solid186 solid elements. At the mating surfaces of the two-stage disc-drum flange and the stop connection, Conta174 contact elements and Targe170 target elements were used to simulate the interface contact state, with a friction coefficient set to 0.15. Bolt preload was applied using Prets179 elements. Bearing and coupling supports were simulated using Combin14 spring-damping elements. Static and modal analyses were performed on the finite element model: In the static analysis, concentrated loads were applied sequentially at each measurement point, and the radial displacement of the corresponding nodes was calculated. The equivalent bending stiffness at each measurement point was obtained based on the force-deformation relationship. In the modal analysis, the natural frequencies and mode shapes of the rotor in the free-free state were calculated, obtaining the first three bending frequencies and their corresponding mode shapes. The simulation results for the first three bending frequencies were compared with the modal test results, with a maximum error of -1.41%. The mode shapes showed good agreement, verifying the accuracy of the established finite element model.

[0027] The modal inversion stiffness method proposed in this invention is applied to a high-pressure rotor of a certain type of aero-engine to verify its engineering applicability. The high-pressure rotor is 1.28 m long and is discretized into 27 beam segments with 49 nodes using beam elements. A stepped shaft is used to simulate a conical shell structure, and 22 stiffness measurement points are defined along the axial direction. The support stiffness at the two support points is set to 3.5 × 10⁻⁶ m. 7 N / m and 1.86×10 7N / m. Modal tests were conducted under free-free boundary conditions, obtaining the first two bending natural frequencies of 335.40 Hz and 567.08 Hz, respectively. Following the procedure described in S2, the elastic modulus and key section parameters were inverted based on the parameterized rotor dynamics model used for static stiffness inversion and the genetic optimization algorithm. The frequencies of the updated parameterized rotor dynamics model under elastic support conditions were compared with the ANSYS simulation results, and the frequency error did not exceed 1.77%. Taking the application of concentrated loads at nodes 13, 26, and 40 as examples, the deflection curves obtained by the self-developed model and the simulation model showed good agreement. The modal inversion stiffness of all 22 measurement points was compared with the ANSYS simulation stiffness, with a maximum error of 4.49%, and the errors of all measurement points were less than 4.5%, meeting the engineering accuracy requirements.

[0028] This invention provides a method for testing the rotor stiffness of an aero-engine based on modal inversion, comprising the following steps: Example 1: Simulated rotor stiffness test of an aero-engine: S1. Obtain the rotor quasi-static stiffness reference and modal parameters.

[0029] In S11, a two-stage disc-drum rotor, including a disc-drum flange, shaft head cone, bearing housing, and drive shaft, is used as a simulated rotor. Figure 2 As shown. The rotor material density is 7850 kg / m³. 3 The Poisson's ratio is 0.3. The required equipment includes: 8 SYNTEK dial gauges (range 0-12.7 mm, accuracy 0.002 mm, resolution 0.001 mm), 1 JHBM-H3 diaphragm force sensor (range 0-5000 N), 1 LMS SCADAS testing system (12 channels), 1 086C01 impact hammer (sensitivity 11.2 mV / N), and 1 B&K-4517 accelerometer (sensitivity 1.032 mV / ms). -2 ).

[0030] In S12, the simulated rotor is rigidly fixed to the T-slot platform, and eight measurement points are defined sequentially along the rotor axis, such as... Figure 2 As shown in the figure. A dial indicator is placed directly below each measurement point to measure the radial displacement at the loading point. The force sensor and dial indicator are zeroed before testing. Loads are applied sequentially to each measurement point, starting from 0 N and increasing in increments of 500 N to 5000 N, while simultaneously recording the displacement data measured by all dial indicators. To evaluate the rotor's circumferential consistency, after completing the initial 0° azimuth loading and measurement, the rotor is rotated sequentially by 90°, 180°, and 270°, and the complete loading and measurement procedure is repeated at each azimuth angle. Taking measurement points 5 and 6 as examples, the displacement-load curves at the four azimuth angles are shown below. Figure 4As shown, the displacement increases monotonically and linearly with the load increasing from 500 N to 4500 N, indicating stable static stiffness. The data points at the four azimuth angles closely overlap and match the fitted stiffness curve well, indicating good circumferential consistency. The stiffness values ​​of each measurement point at each azimuth angle are calculated from the slope of the force-deformation curve, and the average value of the four measurements is taken as the quasi-static stiffness reference value for that measurement point. The average stiffness values ​​of measurement points 5 and 6 are 0.414 × 10⁻⁶ and 0.414 × 10⁻⁶, respectively. 7 N / m and 0.423×10 7 N / m. Following the same procedure, static deformation data were collected and applied sequentially to measurement points 2 through 7. The resulting deflection curves are shown below. Figure 5 As shown, the quasi-static stiffness reference values ​​for each measurement point are detailed in Table 1.

[0031] Table 1. Stiffness values ​​of the simulated rotor at various measuring points based on the quasi-static method.

[0032] In S13, accelerometers are uniformly arranged along the rotor axis, and modal tests with single-point excitation and multi-point vibration pickup are conducted using the hammer excitation method. Data acquisition is performed using the LMS SCADAS testing system, with a sampling frequency of 2048 Hz and a frequency resolution of 0.25 Hz. The frequency response function data of each measuring point are processed using the LMS Test.Lab modal analysis system, and the natural frequencies and mode shapes of the first three bending modes are identified using the least squares complex frequency domain method. The identified first three bending natural frequencies are: first bending 118.75 Hz, second bending 215.77 Hz, and third bending 316.12 Hz, and the corresponding mode shapes are as follows: Figure 5 As shown.

[0033] S2. Establish a parameterized rotor dynamics model for static stiffness inversion. The parameterized rotor dynamics model for static stiffness inversion is as follows: Figure 6 As shown. The rotor is discretized using Timoshenko beam elements, simplifying the shaft segment to beam elements, components such as discs and flanges to lumped mass points, and bearing supports to linear spring-damped elements. The rotor is discretized as... N =15 nodes, including concentrated mass points, shaft segment nodes, and bearing nodes. The nodal displacement vector of the beam element is: in, x i and y i For nodes in x and y Translational displacement in the direction of motion. θ xi and θ yi To bypassx and y Rotational displacement of the shaft. Total number of beam elements. N e =14, the element mass matrix and stiffness matrix are transformed by the matrix... T e After transformation to the global coordinate system, the components are assembled to form the global mass matrix and the global stiffness matrix: The dynamic equations of the rotor system are: in, C Here is the damping matrix. F For external force vectors, q is the nodal displacement vector.

[0034] S3. Construct the modal displacement compliance matrix and establish the optimization objective function based on the modal parameters obtained from the modal tests. When the excitation frequency approaches zero, the displacement frequency response function degenerates into static displacement, thus obtaining the modal displacement compliance matrix: in, For the first i Normalized mode shape of the first order; ω i For the first i The undamped natural frequency of the first order. Discretize the rotor shaft as follows: N For each beam segment, let A be the cross-sectional area of ​​each segment. j ( j =1,2,…, N Let ) be the variable to be optimized, and let f sim,i ( A () is the first eigenvalue obtained by solving the eigenvalues ​​based on the beam model. i First bending frequency, f exp,i The first one measured in the experiment i First bending frequency, construct the optimization objective function: in, δ i For the first i The weighting coefficients for the first frequency are all set to 1; simultaneously, the total rotor mass is used as the weighting factor. M total and cross-sectional area ( A j min Represents the minimum cross-sectional area. A j max (representing the maximum cross-sectional area) is used as a constraint condition, where ρL is the material density. j For the first j The beam segment length is determined to ensure the physical consistency of the model.

[0035] The objective function established in S22 is solved using a genetic optimization algorithm. The iterative process of genetic optimization is as follows: Figure 7 As shown. The algorithm iteratively updates the elastic modulus and cross-sectional area at sensitive locations of the rotor system to minimize the frequency error between the calculated and experimental frequencies. After convergence, the equivalent static compliance matrix is ​​calculated based on the updated model, and the static relationship is then used to determine the compliance matrix. d = Df The inverse analysis outputs the modal inversion stiffness values ​​at each measurement point of the rotor. The comparison between the modal inversion stiffness and the quasi-static test results is shown in Table 2. The maximum relative error between the two methods is -3.14%, indicating that the results obtained by the two methods have good consistency.

[0036] Table 2 Comparison of Modal Inversion Stiffness and Static Stiffness Test Results

[0037] S4. Stiffness Verification and Engineering Applications.

[0038] A three-dimensional solid finite element model consistent with the simulated rotor was established using the finite element software ANSYS, such as... Figure 8 As shown in the figure, the model was meshed using Solid186 solid elements. At the mating surfaces of the two-stage disc-drum flange and the stop connection, Conta174 contact elements and Targe170 target elements were used to simulate the interface contact state, with a friction coefficient set to 0.15. Bolt preload was applied using Prets179 elements. Bearing and coupling supports were simulated using Combin14 spring-damping elements. Static and modal analyses were performed on the finite element model to obtain the simulated stiffness values ​​and the first three bending frequencies at each measurement point. The simulation results for the first three bending frequencies are compared with the modal test results in Table 3. The maximum error is -1.41%, and the mode shapes show good agreement, verifying the accuracy of the established finite element model.

[0039] Table 3 Comparison of natural frequencies between simulation and modal testing

[0040] The quasi-static test stiffness, modal inversion stiffness, and ANSYS simulation stiffness were compared and verified. Taking measurement point 5 and measurement point 6 as examples, the rotor deflection curves obtained by the three methods were plotted, as follows: Figure 9As shown in the figure, the deflection curves obtained by the three methods are all symmetrically parabolic distributed along the axis, and the curve shape, peak position, and amplitude all show good consistency. Table 4 lists the comparison results of stiffness values ​​of the three methods at measurement points 2 to 7. Based on the quasi-static test results, the relative error between the quasi-static test and simulation results is denoted as error 1, and the relative error between the modal inversion stiffness and simulation results is denoted as error 2. Taking measurement point 5 as an example, error 1 is 1.21%, and error 2 is 4.49%; the maximum values ​​of error 1 and error 2 in all measurement points do not exceed 4.5%. The above results show that the modal inversion stiffness method proposed in this invention has good consistency with the quasi-static test and finite element simulation results, verifying the effectiveness of the modal inversion stiffness method.

[0041] Table 4 Comparison of stiffness results obtained by the three methods

[0042] Example 2: Stiffness test of high-pressure rotor of aero-engine: The modal inversion stiffness method proposed in this invention was applied to a high-pressure rotor of a certain type of aero-engine to verify its engineering applicability. This high-pressure rotor is 1.28 m long and is discretized into 27 beam segments with 49 nodes using beam elements. A stepped shaft is used to simulate a conical shell structure, and 22 stiffness measurement points are defined along the axial direction, such as... Figure 10 As shown. The support stiffness at both support points is set to 3.5 × 10⁻⁶. 7 N / m and 1.86×10 7 N / m. Modal tests were conducted under free-free boundary conditions, and the first two bending natural frequencies were obtained as 335.40 Hz and 567.08 Hz, respectively. Following the procedure described in S2 of Example 1, the elastic modulus and key section parameters were inverted based on the parametric rotor dynamics model and genetic optimization algorithm. The frequency of the updated parametric rotor dynamics model under elastic support conditions is compared with the ANSYS simulation results in Table 5. The frequency error does not exceed 1.77%, verifying the accuracy of the updated model.

[0043] Table 5 Frequency Comparison under High-Voltage Rotor Elastic Support Conditions

[0044] Taking the application of concentrated loads at nodes 13, 26, and 40 as an example, the deflection curves obtained from the parameterized rotor dynamics model and the simulation model show good agreement. Figure 11 As shown, the accuracy of the parametric rotor dynamics model in static deformation prediction is verified. Table 6 shows the comparison results of the modal inversion stiffness and ANSYS simulation stiffness at all 22 measurement points. The maximum error is 4.49%, and the errors at all measurement points are less than 4.5%, meeting the engineering accuracy requirements. A comparison of the lateral stiffness distribution of the high-pressure rotor is also provided. Figure 12As shown, the two results agree well. These results demonstrate that the method of the present invention can be effectively applied to the stiffness testing of real, complex aero-engine rotor structures, and has good engineering applicability.

[0045] Table 6 Comparison of Modal Inversion Stiffness and Simulated Stiffness of High-Voltage Rotor

[0046] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0047] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for testing the static stiffness of an aero-engine rotor based on modal inversion, characterized in that, Includes the following steps: Step 1: Obtain the quasi-static stiffness reference and modal parameters of the rotor: Perform quasi-static stiffness test and modal test on the simulated rotor respectively, and obtain the bending stiffness value of the simulated rotor under graded static load as a reference. At the same time, measure the first three bending natural frequencies of the rotor and the corresponding mode shapes. Step 2: Construct a parametric rotor dynamics model for static stiffness inversion. Discretize the rotor along the axial direction into a finite element system containing beam elements, concentrated mass points, and spring-damped supports. Based on the distribution characteristics of the shaft segment, disc, flange, connection interface, and support position in the actual rotor structure, divide the rotor into multiple parameter identification segments corresponding to the sensitivity to stiffness changes, and set equivalent stiffness parameters that can be updated independently for each parameter identification segment. Based on the arrangement of measurement points in the modal test, establish a corresponding mapping relationship between the test measurement points and the nodal degrees of freedom of the parametric rotor dynamics model, so that the calculated bending natural frequencies and mode shapes can be compared with the test modal parameters in a unified coordinate and unified degree of freedom space. Step 3, Genetic Optimization Inversion: Using the bending natural frequency and mode shape obtained in Step 1 as the joint inversion target, the stiffness parameters of each parameter identification section of the parameterized rotor dynamics model in Step 2 are subjected to global iterative optimization with physical constraints. During the optimization process, an objective function is constructed based on the frequency error, mode shape deviation, and coordination relationship of the changes in section parameters between the calculated modal parameters and the experimental modal parameters. Combined with the total rotor mass constraint and the range constraint of the section parameter values, the stiffness distribution of the parameter identification section is updated generation by generation until the objective function meets the preset convergence condition, and the rotor modal inversion stiffness consistent with the measured modal characteristics is obtained.

2. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 1, characterized in that, Step 1 specifically includes the following steps: Step 11: Apply graded static loads radially at different axial positions of the simulated rotor and extract the equivalent bending stiffness values ​​at multiple cross sections of the simulated rotor. Step 12: Perform modal tests on the same simulated rotor, obtain the frequency response function at each measuring point, and identify the first three bending natural frequencies and mode shapes.

3. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 2, characterized in that, In step 11, a graded static load is applied radially at different axial positions of the simulated rotor, with a load range of 0 N ~ 5000 N and a step size of 500 N. The radial deformation at each measurement point is recorded synchronously using a dial indicator to obtain a force-deformation curve. The equivalent bending stiffness value of each section is extracted based on the slope of the force-deformation curve and used as quasi-static stiffness reference data. After completing the initial azimuth angle test, the simulated rotor was rotated 90°, 180° and 270° in sequence, and the loading and measurement were repeated. The average value of the multiple measurement results was taken as the equivalent bending stiffness of the section.

4. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 2, characterized in that, In step 12, multiple effective hammer blows are performed at each excitation point, and the average frequency response function is taken. The frequency response function data of each measurement point are processed by the modal analysis system, and the first three bending natural frequencies and mode shapes are identified by the least squares complex frequency domain method.

5. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 1, characterized in that, Step 2 specifically includes the following steps: The rotor shaft segment was discretized using Timoshenko beam elements, simplifying the shaft segment into beam elements, the disc and flange into concentrated mass points, and the bearing support into linear spring-damped elements. Based on the rotor geometry, abrupt changes in cross-section, disc-shaft connection, flange connection, and support positions, the rotor was divided into several parameter identification segments along the axial direction. Each parameter identification segment was assigned independent cross-sectional parameters and / or equivalent material parameters to form a segmented parameterized model that could characterize the non-uniform distribution of local stiffness in the rotor. Based on the sensor placement and excitation positions in the modal test, a spatial correspondence between the test measurement points and the nodes of the finite element model was established. The modal shapes obtained from the test were mapped to the corresponding node degrees of freedom of the finite element model after interpolation, normalization, and degree-of-freedom transformation to ensure that the test modal parameters and the calculated modal parameters were calculated for error on the same evaluation basis. Rotor discretization N Each node constructs a mass matrix. M Stiffness matrix K and damping matrix C The dynamic equations of the rotor system: in, F For external force vectors, q is the nodal displacement vector.

6. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 1, characterized in that, Step 3 specifically includes: Step 31, Variable Definition: Define the cross-sectional area of ​​each sensitive shaft segment of the rotor. A j The elastic modulus is set as the optimization variable, and the optimization variable is used to characterize the local bending stiffness distribution of the rotor in different sections along the axial direction. The optimization variables of each parameter identification section are set with independent upper and lower limits according to the corresponding structural position to reflect the differences in actual structural size and material properties of different sections. Step 32, Objective Function Construction: Define and compute the objective function. F ( A The weighted sum of squares of the relative errors between the simulated natural frequencies and the experimental natural frequencies is: in, δ i For the first i The weighting coefficients for the first-order frequencies are all set to 1; simultaneously, the total rotor mass and cross-sectional area are used as constraints, where... ρ For material density, L j For the first j The beam segment length is determined to ensure physical consistency of the model; the rotor shaft is discretized into... N Each beam segment, f sim,i ( A () is the first eigenvalue obtained by solving the eigenvalues ​​based on the beam model. i First bending frequency, f exp,i The first one measured in the experiment i First bending frequency; Step 33, Optimization: A genetic algorithm is used to perform a global search on the optimization variables. The parameter vector is updated generation by generation through selection, crossover, and mutation operations. In each iteration, each candidate parameter vector is substituted into the parameterized dynamic model established in Step 2 to solve for eigenvalues, obtain the corresponding bending natural frequencies and mode shapes, and calculate the matching degree of the candidate parameter vectors with the test modal parameters based on the objective function. Candidate parameter vectors that meet the constraints and have a better objective function are retained and evolved until the objective function reaches the preset minimum tolerance value, or the rate of change of the optimal objective function is less than a set threshold for several consecutive generations. Then, the iteration stops and the current optimal parameter vector is output as the identification result of the rotor stiffness distribution parameters, and the modal inversion stiffness is obtained accordingly.

7. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 6, characterized in that, The specific process of step 33 includes: Step 331: Initialize the population: Within the preset physical feasible domain of the cross sections and elastic moduli of each shaft segment of the rotor system, randomly generate P initial parameter vector populations that satisfy the above mass constraints. Step 332, Fitness Assessment: Substitute the parameter vector of each individual in the population into the parameterized dynamics model to solve for the eigenvalues ​​and obtain the first... i 1st-order simulated bending natural frequency f sim,i And calculate the corresponding objective function value. F ( A ); 1 / (1+ F ( A The fitness value is defined as the individual's fitness value, used to quantify how closely the individual approximates the experimental data. Step 333, Genetic Operator Operations: Based on fitness values, perform selection operations on individuals in the population, retaining individuals with high fitness; perform crossover and mutation operations on the selected samples to generate a new generation of parameter vector population, and perform quality constraint verification on the new generation population, removing individuals that do not meet the constraints; Step 334, Iterative Convergence Determination: Repeat steps 332 to 333 until the... k Objective function of generation population F ( A When the preset minimum value is reached, or when the change in the objective function value for several consecutive generations is less than the set relative error threshold, the iteration stops and the current optimal parameter vector is output as the identification result of the stiffness distribution parameter.

8. The method for testing the static stiffness of an aero-engine rotor based on modal inversion according to claim 6, characterized in that, The modal inversion stiffness calculation process includes: Based on the bending natural frequencies and mode shapes identified in step 3, when the excitation frequency approaches zero, the displacement frequency response function degenerates into a static displacement response, thereby establishing the rotor's modal displacement compliance matrix: in, For the first i Normalized mode shape of the first order; ω i For the first i The undamped natural frequency of each order is determined; the modal parameters obtained from each order identification are substituted into the modal displacement compliance matrix to obtain the equivalent static compliance matrix corresponding to the position of each measuring point or target section; and by inverting the equivalent static compliance matrix, the mapping from dynamic modal characteristics to static stiffness values ​​is realized, thereby obtaining the modal inversion stiffness of the rotor.