A method and system for online identification of rotor modes of a magnetic levitation motor

Abstract

This invention discloses an online mode identification method and system for a magnetic levitation motor rotor. The method includes: obtaining the rotor simulation mode frequency based on rotor simulation results; setting an excitation frequency range centered on the rotor simulation mode frequency while the rotor is in a levitation state, and applying an adjustable sinusoidal excitation signal at the controller output position; during the excitation process, acquiring displacement signals in the X and Y directions and power amplifier current signals along the rotor radial direction; calculating the phase difference between the displacement and the power amplifier current signal in each direction; within the excitation frequency range, the excitation frequency corresponding to the sudden change in phase difference is the rotor's current order mode frequency; obtaining the higher-order mode frequencies of the rotor by frequency sweeping to complete the identification; this invention does not require additional dedicated excitation equipment and can identify mode frequencies and vibration modes online under different operating states such as rotor levitation, low-speed rotation, and high-speed rotation, and has the advantages of simple implementation, high identification accuracy, and no impact on the normal operation of the system.

Description

Technical Field

[0001] This invention relates to the field of magnetic levitation rotating machinery technology, specifically to an online identification method and system for the rotor modes of a magnetic levitation motor. Background Technology

[0002] A magnetic levitation motor is a motor whose rotor is suspended by a magnetic bearing. With the adoption of magnetic bearing support technology, the rotor of the motor and the bearing are no longer in contact. It has been increasingly widely used in the field of magnetic levitation fluid machinery such as blowers, air compressors, vacuum pumps, refrigeration compressors, and organic Rankine cycles.

[0003] High-speed permanent magnet motor rotors are typically elongated and slender, significantly increasing their likelihood of approaching critical speeds compared to conventional speed motors. When the motor's rated speed approaches the critical speed, the rotor will experience severe vibrations, potentially leading to rotor damage and jeopardizing the motor's safe operation. Therefore, accurate analysis of the modal characteristics of high-speed permanent magnet motor rotors is crucial.

[0004] Current technology first uses UG software to build a rotor model, then uses finite element analysis software to evaluate its modes and obtain the rotor's natural frequencies. Experimentally, the natural frequencies of the rotor are determined using the LMS impact test technique, achieving a comparison and verification between simulation and experiment. However, determining the rotor's natural frequencies using the LMS impact test technique has the following problems and drawbacks.

[0005] 1) The force, angle, and position of manual tapping are unstable, which can easily cause local plastic deformation and contact nonlinearity, resulting in distortion of the frequency response function and low coherence coefficient; 2) The hammer pulse energy is concentrated in the low frequency, and the high frequency excitation is insufficient. The high-speed rotor's high-order modes (>5kHz) are often not fully excited, the signal-to-noise ratio is low, and mode truncation and frequency leakage occur. 3) The accelerometer is attached to the rotor surface. For lightweight, high-speed rotors, this will reduce the natural frequency and change the mode shape, introducing measurement errors. 4) The sparse arrangement of measuring points and insufficient circumferential measuring points lead to incomplete vibration mode identification and difficulty in capturing complex vibration modes such as ellipticization and local bending. 5) In scenarios with low damping and dense modes, the steady-state simulation algorithm and PolyMAX algorithm of LMS software are prone to frequency aliasing and poor mode orthogonality. 6) High cost: The basic starting price for a set of LMS hammering equipment is around 250,000 yuan; 7) Online measurement is not possible; the rotor must be removed from the machine for measurement, which is time-consuming and labor-intensive. 8) It cannot reproduce the hot state condition; it can only be measured in the cold state. 9) The test mainly focuses on bending modes, with low accuracy in torsional mode testing, incomplete rotor mode identification, and additional mass interference. 10) It can only measure natural frequency and mode shape, but cannot test rotational effect and magnetic bearing related parameters. Summary of the Invention

[0006] Technical objective: To address the shortcomings of existing magnetic levitation motor rotor mode identification methods, this invention discloses an online method and system for magnetic levitation motor rotor mode identification.

[0007] Technical solution: To achieve the above technical objectives, the present invention adopts the following technical solution: A method for online identification of rotor modes of a magnetic levitation motor includes the following steps: S01. Based on the rotor simulation results, obtain the rotor simulation modal frequencies; S02. In the rotor suspension state, an adjustable sinusoidal excitation signal is applied to the output position of the rotor controller. The excitation frequency range of the sinusoidal excitation signal is set with the rotor simulation mode frequency as the center value. S03. While applying a sinusoidal excitation signal, the displacement signals and power amplifier current signals in the X and Y directions are collected along the rotor radial direction. The X and Y directions are two directions on the rotor radial plane that form a 90° angle with each other.

[0008] S04. Calculate the difference between the phase of the rotor X-direction displacement signal and the phase of the corresponding power amplifier current signal, and the difference between the phase of the Y-direction displacement signal and the phase of the corresponding power amplifier current signal. S05. Within the frequency range of the sinusoidal excitation signal, the excitation frequency corresponding to the phase difference abrupt change is the rotor's modal frequency at the current order. The higher-order modal frequencies of the rotor are obtained by frequency sweeping to complete rotor mode identification.

[0009] Preferably, the sinusoidal excitation signal of the present invention is Where A is the amplitude of the sinusoidal excitation signal, and ω is the excitation angular frequency. f is the excitation frequency, and t is the control period; the excitation frequency range is the simulation modal frequency ±50Hz, and the sinusoidal excitation signal is output in increments of 1 Hz.

[0010] Preferably, in step S03 of the present invention, the acquisition frequency of the displacement signal and the power amplifier current signal is greater than twice the upper limit of the excitation frequency.

[0011] Preferably, in step S04 of the present invention, the process of calculating the phase includes: S041. Using the sinusoidal excitation signal as a reference, generate two reference signals with the same frequency and orthogonal phase, denoted as the in-phase reference signal I_ref=sin(ωt) and the orthogonal reference signal Q_ref=cos(ωt); S042. The response signal corresponding to the rotor displacement or power amplifier current is expressed as x(t)=B·sin(ωt+φ), where B is the amplitude of the response signal and φ is the phase of the response signal. S043. Multiply the response signal with the in-phase reference signal and the quadrature reference signal respectively, and filter out the high-frequency components in the signal by passing it through a low-pass filter with a cutoff frequency range of 0.1ω-0.2ω, retaining the DC component, to obtain the real part I and the imaginary part Q of the component in the response signal that is in the same frequency as the reference signal; S044. Based on the geometric relationship between the real and imaginary parts, obtain the amplitude of the corresponding response signal. phase .

[0012] Preferably, in step S043 of the present invention, the corresponding DC component is calibrated by multiplying it by a coefficient 2 to obtain the accurate real part I and imaginary part Q.

[0013] Preferably, the sinusoidal excitation signal and the reference signal of the present invention are both generated by a DSP, and the calculation of the amplitude and phase of the response signal is realized by the arithmetic operation module inside the DSP.

[0014] Preferably, when performing rotor mode identification, the present invention simultaneously acquires displacement signals at both ends of the rotor, identifies the mode shape of the rotor based on the amplitude distribution of the displacement signals at both ends of the rotor, and locates the mode shape nodes and antinodes.

[0015] Preferably, in step S02 of the present invention, under the rotor suspension state, the response signals of the rotor to the sinusoidal excitation signal are tested respectively under the conditions of rotor suspension, rotor low speed rotation, and rotor high speed rotation, so as to obtain the modal parameters of the rotor under different operating conditions.

[0016] This invention discloses an online rotor mode identification system for a magnetic levitation motor, which uses the above-mentioned identification method to identify rotor modes.

[0017] Beneficial effects: The online mode identification method and system for magnetic levitation motor rotor disclosed in this invention has the following beneficial effects: 1. This invention, by suspending the rotor and then adding a sinusoidal excitation signal with a variable excitation frequency generated by a DSP to the controller output, identifies the modal frequencies of the magnetically levitated rotor by measuring the phase difference of the acquired response signals. In engineering practice, this achieves real-time online identification of rotor modes for high-speed magnetically levitated rotors. This invention leverages the advantages of active and controllable magnetic levitation, eliminating the need for additional testing equipment. It is low-cost, simple in structure, requires minimal computational resources, and offers strong real-time performance. It can obtain the modal frequencies of cold and hot magnetically levitated rotors online, solving the problem of requiring rotor disassembly and equipment processing for modal testing of magnetically levitated rotating machinery systems.

[0018] 2. The test conditions of this invention are more realistic: the rotor is in a suspended state, which is close to the actual operating conditions of a high-speed magnetic levitation motor, and can be used to test the rotor rotation; the LMS hammer impact method can only test a stationary rotor, which is out of touch with the actual operating conditions.

[0019] 3. The testing process of this invention involves no mechanical contact, no additional mass interference, and no risk of local plastic deformation; the LMS hammering method is a mechanical contact excitation, which is prone to introducing contact nonlinearity and additional errors.

[0020] 4. More comprehensive and accurate frequency identification: The frequency sweep covers the entire frequency band and can effectively identify high-order mode frequencies without frequency leakage. The front and rear radial dual-path frequency sweep can fully capture the rotor bending characteristics; the LMS hammer method has insufficient high-frequency excitation, and high-order modes are prone to frequency leakage and truncation.

[0021] 5. In addition to testing the rotor's natural frequency and mode shape, this invention can also test the rotor's torsional / rigid modes, damping ratio, frequency response function, magnetic bearing stiffness / damping, and modal parameters under rotational effects; the LMS hammer impact method can only test the natural frequency and mode shape, and cannot test rotational effects and related parameters of magnetic bearings.

[0022] 6. This invention can accurately locate vibration mode nodes and antinodes by synchronously acquiring data from two measuring points at both ends of the rotor, without any additional interference from sensors; the LMS hammer impact method has sparse measuring points, measurement blind spots, and is easily distorted in vibration mode identification due to the influence of the added mass of the accelerometer.

[0023] 7. The errors of this invention mainly come from electromagnetic noise and suspension stability, and the interference can be suppressed by means of filtering; the LMS hammering method is greatly affected by the impact stability, added mass and environmental vibration, and the repeatability error is obvious.

[0024] 8. The identification method of the present invention is specifically designed for magnetic levitation motor rotors and can realize modal testing in high-speed rotation scenarios; the LMS hammer impact method is a general rotor testing method and is not suitable for levitation state and high-speed operating conditions. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0026] Figure 1 This is a schematic diagram illustrating the application of a sinusoidal excitation signal to this invention. Detailed Implementation

[0027] Reference will now be made in detail to embodiments of the present disclosure, one or more of which are set forth herein. Each embodiment and example is provided by way of explanation of the apparatus, composition, and materials of the present disclosure, and not by way of limitation. Rather, the following description provides convenient illustrations for implementing exemplary embodiments of the present disclosure. Indeed, it will be apparent to those skilled in the art that various modifications and variations can be made to the teachings of the present disclosure without departing from the scope or spirit of the present disclosure.

[0028] A method for online identification of rotor modes of a magnetic levitation motor includes the following steps: S01. Based on the rotor simulation results, obtain the rotor simulation modal frequencies; First, a rotor model is built using UG software. Then, finite element analysis software is used to simulate and evaluate its modes to obtain the rotor's natural frequencies. These natural frequencies are the simulated modal frequencies, used to provide a benchmark for testing, in order to define the test frequency range and improve efficiency. For mature products, if the approximate range of natural frequencies is known, this step can be omitted.

[0029] S02, In the rotor suspension state, such as Figure 1 As shown, an adjustable sinusoidal excitation signal is applied at the output position of the rotor controller. The excitation frequency range of the sinusoidal excitation signal is set with the rotor simulation modal frequency as the center value.

[0030] Figure 1 In the diagram, point a represents the reference displacement signal, corresponding to the rotor's suspension center; point g represents the output of the displacement sensor, and q is the excitation signal; point b is the controller's input, and point c represents the controller's output; point d represents the power amplifier's input, and point e represents the power amplifier's output. The power amplifier's current signal, i.e., the magnetic bearing current, is detected by the current sensor, and point f represents the electromagnetic force output from the magnetic bearing to the rotor. The rotor's displacement response is ultimately reflected in the change in the displacement sensor's output.

[0031] The sinusoidal excitation signal is Where A is the amplitude of the sinusoidal excitation signal, and ω is the excitation angular frequency. f is the excitation frequency, t is the control period, and Pi is the mathematical constant pi, with a value of 3.14. The excitation frequency range is the simulation modal frequency ±50Hz. The sinusoidal excitation signal is output in increments of 1 Hz, and the rotor test is performed by gradually sweeping the frequency.

[0032] In the rotor suspension state, the response signals of the rotor to the sinusoidal excitation signal can be tested by changing the rotor state, respectively, when the rotor is only suspended, rotating at low speed, and rotating at high speed, so as to obtain the modal parameters of the rotor under different operating conditions.

[0033] S03. While applying a sinusoidal excitation signal, the displacement signals and power amplifier current signals in the X and Y directions are acquired along the rotor radial direction. The X and Y directions are two directions on the rotor radial plane that form a 90° angle with each other. For example, the X direction is horizontal and the Y direction is vertical, or the X direction is at a 45° angle with the horizontal or vertical direction as a reference, and the Y direction is at a 135° angle with the same reference. The acquisition frequency of the displacement signal and the power amplifier current signal is greater than twice the upper limit of the excitation frequency to avoid frequency aliasing, thereby improving the accuracy of the acquired data.

[0034] S04. Calculate the difference between the phase of the rotor X-direction displacement signal and the phase of the corresponding power amplifier current signal, and the difference between the phase of the Y-direction displacement signal and the phase of the corresponding power amplifier current signal. The difference is the phase difference.

[0035] The process of calculating the phase includes: S041. Using the sinusoidal excitation signal as a reference, generate two reference signals with the same frequency and orthogonal phase, denoted as the in-phase reference signal I_ref=sin(ωt) and the orthogonal reference signal Q_ref=cos(ωt); S042. The response signal corresponding to the rotor displacement or power amplifier current is expressed as x(t)=B·sin(ωt+φ), where B is the amplitude of the response signal and φ is the phase of the response signal. S043. Multiply the response signal with the in-phase reference signal and the quadrature reference signal respectively, and filter out the high-frequency components in the signal by passing it through a low-pass filter with a cutoff frequency range of 0.1ω-0.2ω, retaining the DC component, to obtain the real part I and the imaginary part Q of the component in the response signal that is in the same frequency as the reference signal; S044. Based on the geometric relationship between the real and imaginary parts, obtain the amplitude of the corresponding response signal. phase .

[0036] Through the above calculation process, the phase of the rotor displacement signals in the X and Y directions and the corresponding power amplifier current signals can be obtained respectively, thereby calculating the phase difference.

[0037] To further improve testing accuracy, in step S043 of this invention, the corresponding DC component is multiplied by a coefficient of 2 for calibration to obtain accurate real part I and imaginary part Q. In this invention, both the sinusoidal excitation signal and the reference signal are generated by a DSP. The calculation of the amplitude and phase of the response signal is implemented through the arithmetic operation module inside the DSP, eliminating the need for complex frequency domain conversion and enabling real-time output of amplitude and phase values, thus adapting to the dynamic testing requirements during frequency sweeping.

[0038] S05. Within the frequency range of the sinusoidal excitation signal, the excitation frequency corresponding to the phase difference mutation is the rotor's modal frequency at the current order. The higher-order modal frequencies of the rotor are obtained by frequency sweeping to complete rotor mode identification. The angle of the phase difference mutation is generally 90° or fluctuates within a small range centered on 90°.

[0039] When the excitation frequency is not equal to the rotor's natural frequency, the displacement and the power amplifier current response are at the same frequency and the phase difference is stable. When the excitation frequency is close to or equal to the rotor's bending natural frequency, the rotor enters a resonance state, and the system stiffness and damping characteristics change significantly, resulting in a significant abrupt change in the phase difference between the displacement signal and the power amplifier current signal. The excitation frequency corresponding to the abrupt change in phase difference is the natural frequency of the rotor's corresponding order.

[0040] This invention utilizes the abrupt change in the phase difference between the displacement signal and the power amplifier current signal to identify modal frequencies. It can test the rotor bending mode frequency, torsional mode frequency, and rigid mode frequency, making up for the shortcomings of the LMS hammer impact method in high-order mode leakage and low accuracy of torsional mode testing. By changing the excitation frequency, the phase difference and amplitude variation law between the displacement signal and the power amplifier current signal are analyzed, and the curves of the change of magnetic bearing stiffness and damping with frequency are inversely derived, making up for the errors caused by the simplification of magnetic bearings in the simulation.

[0041] This invention, when performing rotor mode identification, can simultaneously acquire displacement signals at both ends of the rotor. Based on the amplitude distribution of the displacement signals at both ends of the rotor, it identifies the mode shapes of the rotor and locates the mode shape nodes and antinodes, such as the radial vibration distribution of the bending mode and the circumferential rotation characteristics of the torsional mode. By analyzing the resonance peak width of the "excitation frequency-displacement / current response amplitude" curve and combining it with the calculated signal amplitude data, the damping ratio of each rotor mode (including material damping, structural damping, and active damping of the magnetic bearing) can be calculated, reflecting the rotor's vibration attenuation capability and providing a basis for optimizing the magnetic bearing control parameters (PID).

[0042] Simultaneously, by synchronously acquiring excitation signals, displacement signals, and power amplifier current response signals through DSP, a "frequency-amplitude / phase" frequency response function can be generated, which clearly reflects the vibration response characteristics of the rotor at different frequencies. This data is the core data for simulation and experimental comparison and error tracing, and can also be used to evaluate the frequency dependence of magnetic bearing stiffness and damping.

[0043] The present invention also discloses an online identification system for rotor modes of a magnetic levitation motor, which uses the above-mentioned identification method to identify rotor modes.

Claims

1. A method for online identification of rotor modes of a magnetic levitation motor, characterized in that, Including the following steps: S01. Based on the rotor simulation results, obtain the rotor simulation modal frequencies; S02. In the rotor suspension state, an adjustable sinusoidal excitation signal is applied to the output position of the rotor controller. The excitation frequency range of the sinusoidal excitation signal is set with the rotor simulation mode frequency as the center value. S03. While applying a sinusoidal excitation signal, the displacement signals and power amplifier current signals in the X and Y directions are collected along the rotor radial direction. The X and Y directions are two directions on the rotor radial plane that form a 90° angle with each other. S04. Calculate the difference between the phase of the rotor X-direction displacement signal and the phase of the corresponding power amplifier current signal, and the difference between the phase of the Y-direction displacement signal and the phase of the corresponding power amplifier current signal. S05. Within the frequency range of the sinusoidal excitation signal, the excitation frequency corresponding to the phase difference abrupt change is the rotor's modal frequency at the current order. The higher-order modal frequencies of the rotor are obtained by frequency sweeping to complete rotor mode identification.

2. The method for online identification of rotor modes of a magnetic levitation motor according to claim 1, characterized in that, The sinusoidal excitation signal is Where A is the amplitude of the sinusoidal excitation signal, and ω is the excitation angular frequency. f is the excitation frequency, and t is the control period; the excitation frequency range is the simulation modal frequency ±50Hz, and the sinusoidal excitation signal is output in increments of 1 Hz.

3. The method for online identification of rotor modes of a magnetic levitation motor according to claim 1, characterized in that, In step S03, the acquisition frequency of the displacement signal and the power amplifier current signal is greater than twice the upper limit of the excitation frequency.

4. The method for online identification of rotor modes of a magnetic levitation motor according to claim 1, characterized in that, In step S04, the process of calculating the phase includes: S041. Using the sinusoidal excitation signal as a reference, generate two reference signals with the same frequency and orthogonal phase, denoted as the in-phase reference signal I_ref=sin(ωt) and the orthogonal reference signal Q_ref=cos(ωt); S042. The response signal corresponding to the rotor displacement or power amplifier current is expressed as x(t)=B·sin(ωt+φ), where B is the amplitude of the response signal and φ is the phase of the response signal. S043. Multiply the response signal with the in-phase reference signal and the quadrature reference signal respectively, and filter out the high-frequency components in the signal by passing it through a low-pass filter with a cutoff frequency range of 0.1ω-0.2ω, retaining the DC component, to obtain the real part I and the imaginary part Q of the component in the response signal that is in the same frequency as the reference signal; S044. Based on the geometric relationship between the real and imaginary parts, obtain the amplitude of the corresponding response signal. phase .

5. The method for online identification of rotor modes of a magnetic levitation motor according to claim 4, characterized in that, In step S043, the corresponding DC component is calibrated by multiplying it by a coefficient of 2 to obtain the accurate real part I and imaginary part Q.

6. The method for online identification of rotor modes of a magnetic levitation motor according to claim 4, characterized in that, Both the sinusoidal excitation signal and the reference signal are generated by a DSP, and the calculation of the amplitude and phase of the response signal is implemented by the arithmetic operation module inside the DSP.

7. The method for online identification of rotor modes of a magnetic levitation motor according to claim 6, characterized in that, During rotor mode identification, displacement signals at both ends of the rotor are collected simultaneously. Based on the amplitude distribution of the displacement signals at both ends of the rotor, the mode shape of the rotor is identified, and the positions of the mode shape nodes and antinodes are located.

8. The method for online identification of rotor modes of a magnetic levitation motor according to claim 1, characterized in that, In step S02, under the rotor suspension state, the response signals of the rotor to the sinusoidal excitation signal are tested under the following conditions: rotor suspension, rotor low speed rotation, and rotor high speed rotation, respectively, to obtain the modal parameters of the rotor under different operating conditions.

9. A magnetic levitation motor rotor mode online identification system, characterized in that, Rotor mode identification is performed using the identification method described in any one of claims 1-8.

Images