Magnetic suspension bearing high-frequency vibration suppression method and system based on carrier phase-shifting PWM (Pulse Width Modulation)

Abstract

The invention discloses a magnetic suspension bearing high-frequency vibration suppression method and system based on carrier phase-shifting PWM. The method comprises the steps that the carrier phase-shifting angle of each driving winding of a magnetic bearing is regulated and controlled; the switching frequency of PWM control is adjusted, and a vibration response minimum point is found; or the carrier phase shift angle of each driving winding of the magnetic bearing is regulated and controlled, the switching frequency of PWM control is regulated, and the vibration response minimum point is found. The method comprises the following steps: designing an experiment, collecting vibration data under different phase shift angles and switching frequencies, training a deep neural network prediction model, establishing a mapping relation between PWM regulation and control parameters and vibration magnitude, and further performing joint optimization on the PWM regulation and control parameters (phase shift angles and switching frequencies of carrier waves of windings). By regulating and controlling the carrier phase-shifting angle and the switching frequency, the vibration of the high-frequency switching frequency of the magnetic bearing is reduced by more than 80%, the current ripple amplitude of a direct-current bus is reduced by 80%, and meanwhile, the harmonic current of a magnetic bearing winding and the rotor displacement control precision are not influenced.

Description

Technical Field

[0001] The present invention belongs to the intersection field of artificial intelligence and magnetic suspension system, and more specifically, relates to a method and system for suppressing high-frequency vibration of magnetic suspension bearings based on carrier phase-shift PWM. Background Art

[0002] Ships play an important role in the military field. With the development of the national defense industry, higher requirements are placed on the stealth of ships. However, due to the lack of active control methods, the vibration problem caused by rotating mechanical equipment supported by traditional sliding bearings is prominent and difficult to change. Magnetic bearing (magnetic bearing) technology realizes contactless support function by actively controlling the strength of the magnetic field. On the one hand, it realizes contactless support and greatly reduces the mechanical vibration caused by bearing friction. On the other hand, it has active regulation capabilities, which provides more possibilities for optimizing the electromagnetic vibration level.

[0003] However, since the magnetic bearings are powered by PWM (Pulse-width modulation), harmonic currents related to the switching frequency will be introduced into the windings. The harmonic currents excite high-frequency electromagnetic forces in the air gap and act on the stator, ultimately leading to vibration noise at the switching frequency, deteriorating the quietness of the electrical equipment.

[0004] In order to suppress the switching frequency vibration noise caused by PWM, existing research often focuses on the motor inverter side, and there is little research on the high-frequency vibration introduced by the magnetic bearing controller. However, through the analysis of the vibration spectrum of the magnetic bearing motor through vibration instruments, it is found that the high-frequency vibration level introduced by the magnetic bearing controller is no less than the high-frequency vibration introduced by the motor inverter. Therefore, it is very necessary to control the high-frequency vibration characteristic line spectrum of the magnetic bearing.

[0005] The carrier phase-shift PWM technology was first used to improve the common-mode voltage waveform by changing the phase difference between different carriers. In recent years, some domestic and foreign research institutions have also used it to suppress high-frequency vibrations of multi-phase or multi-module motors. However, all research is based on motors with traditional sliding bearings, so carrier phase shifting is currently only used in the control of motor inverters. In addition, although the existing carrier phase-shift PWM on the motor side can reduce high-frequency vibrations, it will cause an increase in harmonic currents in the motor windings, thereby reducing system efficiency. In addition, existing carrier phase-shifting technology often only considers changes in the carrier phase and ignores the impact of the switching frequency. In fact, it was found in research that the switching frequency plays an important role in the effect of carrier phase shifting. Therefore, the combined optimization of PWM control parameters such as carrier phase and frequency is expected to deeply optimize the high-frequency vibration level of magnetic bearings. Summary of the invention

[0006] In view of the defects of the prior art, the purpose of the present invention is to provide a method and system for suppressing high-frequency vibration of magnetic bearings based on carrier phase-shift PWM, which aims to reduce the switching frequency vibration of the magnetic bearings without affecting the normal operating performance of the system, and optimize the values ​​of various PWM control parameters through artificial intelligence methods to achieve a significant suppression effect of high-frequency vibration.

[0007] To achieve the above object, the present invention provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, wherein the high-frequency vibration to be suppressed is the vibration of the motor housing (body) or the motor foot mainly at the switching frequency and its multiples, including:

[0008] A data set of phase shift angle and high-frequency vibration of the carrier phase is used to construct a machine learning model, wherein the machine learning model is used to fit the nonlinear relationship between the phase shift angle and the high-frequency vibration, and the trained machine learning model is used as a high-frequency vibration prediction model for obtaining the phase shift angle in a continuously changing range;

[0009] Carrier phase shifting is performed on each driving winding of the magnetic suspension bearing respectively; wherein, for a magnetic suspension bearing having n sets of driving windings, the phase shifting method includes: taking the carrier phase of the nth set of driving windings as a reference, the 1st, 2nd ... (n-1)th sets of windings are shifted relative to the carrier phase of the nth set of driving windings, and the phase shift angles are θ1, θ2 ... θ n-1 ;

[0010] Taking minimizing high-frequency vibration as the optimization goal, the optimal phase shift angle combination (θ1, θ2...θ n-1 ), thereby achieving high-frequency vibration suppression of the magnetic bearing.

[0011] Furthermore, the phase shift angles θ1, θ2, ... n-1 The range is [0,360).

[0012] Furthermore, the optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

[0013] The present invention also provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, comprising:

[0014] A machine learning model is constructed using a data set of PWM switching frequency and high-frequency vibration. The machine learning model is used to fit the nonlinear relationship between the PWM switching frequency and the high-frequency vibration. The trained machine learning model is used as a high-frequency vibration prediction model for obtaining a phase shift angle in a continuously changing range.

[0015] The PWM switching frequency f of the magnetic bearing is adjusted, and the adjustment method includes increasing or decreasing f;

[0016] Taking minimizing high-frequency vibration as the optimization goal, the optimal PWM switching frequency f is found through the high-frequency vibration prediction model, so as to achieve high-frequency vibration suppression of magnetic suspension bearings.

[0017] Furthermore, the optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

[0018] The present invention also provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, comprising:

[0019] A data set of phase shift angle, PWM switching frequency and high-frequency vibration of the carrier phase is used to construct a machine learning model, wherein the machine learning model is used to fit the nonlinear relationship between the phase shift angle, PWM switching frequency and high-frequency vibration, and the trained machine learning model is used as a high-frequency vibration prediction model for obtaining the phase shift angle and PWM switching frequency in a continuously changing range;

[0020] Carrier phase shifting is performed on each driving winding of the magnetic suspension bearing respectively; wherein, for a magnetic suspension bearing having n sets of driving windings, the phase shifting method includes: taking the carrier phase of the nth set of driving windings as a reference, the 1st, 2nd ... (n-1)th sets of windings are shifted relative to the carrier phase of the nth set of driving windings, and the phase shift angles are θ1, θ2 ... θ n-1 ; Adjust the PWM switching frequency f of the magnetic bearing, and the adjustment method includes increasing or decreasing f;

[0021] Taking minimizing high-frequency vibration as the optimization goal, the optimal phase shift angle and PWM switching frequency combination (θ1, θ2...θ n-1 , f), thereby achieving high-frequency vibration suppression of the magnetic bearing.

[0022] Furthermore, the optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

[0023] The present invention also provides a high-frequency vibration suppression system for magnetic suspension bearings based on carrier phase-shift PWM, comprising: a computer-readable storage medium and a processor;

[0024] The computer-readable storage medium is used to store executable instructions;

[0025] The processor is used to read the executable instructions stored in the computer-readable storage medium to execute any of the above-mentioned methods for suppressing high-frequency vibration of magnetic suspension bearings based on carrier phase-shift PWM.

[0026] Compared with the prior art, the above technical solutions conceived by the present invention can achieve the following

[0027] Beneficial effects:

[0028] 1. The present invention provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM. Due to the relative independence of the winding structure of the magnetic suspension bearing, the application of carrier phase-shift PWM technology will not cause changes in the harmonic current of the winding, but may reduce the harmonic current on the DC bus side and improve the system efficiency. In order to suppress high-frequency vibration, three means are adopted in the carrier phase-shift PWM technology: a single carrier phase-shift method, a single frequency change method, or both methods are performed simultaneously. Among them, the single carrier phase-shift method uses a machine model to fit the nonlinear relationship between the phase shift angle and high-frequency vibration, and finds the optimal phase shift angle combination through a high-frequency vibration prediction model, thereby changing the spatial form of the high-frequency electromagnetic force in the air gap, exciting different vibration membrane states, and suppressing high-frequency vibration; the single frequency change method uses a machine model to fit the nonlinear relationship between the PWM switching frequency and high-frequency vibration, and finds the optimal carrier switching frequency through a high-frequency vibration prediction model, and finds the lowest point of the vibration response; the present invention also innovatively integrates carrier phase shift and switching frequency adjustment into carrier phase-shift PWM control, which can deeply optimize the high-frequency vibration of the switching frequency, and the amplitude can be reduced by more than 80% under both static suspension and rotation conditions.

[0029] 2. Compared with the traditional motor controller high-frequency vibration suppression method, which establishes a high-frequency vibration prediction model through analytical solution or numerical simulation, there are problems such as insufficient modeling accuracy, cumbersome and time-consuming, and difficulty in determining the optimized phase angle. The present invention proposes a high-frequency vibration modeling method based on machine learning, which establishes the relationship between various PWM control parameters and high-frequency vibration through experimental data drive, can more accurately describe the actual vibration problem, and find the optimal carrier phase shift angle and switching frequency. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A schematic diagram of the structure of a magnetic bearing provided in an embodiment of the present invention.

[0031] Figure 2 A schematic diagram of a carrier phase-shift PWM provided in an embodiment of the present invention.

[0032] Figure 3 A schematic diagram of the change of high-frequency electromagnetic force with the carrier phase shift angle and the vibration mode change excited by it provided in an embodiment of the present invention.

[0033] Figure 4 The neural network prediction results of the magnetic bearing suspension working condition provided in the embodiment of the present invention are compared with the displacement sensor measurement results.

[0034] Figure 5 A schematic diagram of the modal superposition theorem and mechanical transfer function provided in an embodiment of the present invention.

[0035] Figure 6Comparison of high frequency vibration results before and after carrier phase shifting provided by an embodiment of the present invention.

[0036] Figure 7 The embodiment of the present invention provides an evaluation of system operating performance before and after carrier phase shifting. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0038] The present invention provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, comprising:

[0039] A data set of phase shift angle and high-frequency vibration of the carrier phase is used to construct a machine learning model, wherein the machine learning model is used to fit the nonlinear relationship between the phase shift angle and the high-frequency vibration, and the trained machine learning model is used as a high-frequency vibration prediction model for obtaining the phase shift angle in a continuously changing range;

[0040] Carrier phase shifting is performed on each driving winding of the magnetic suspension bearing respectively; wherein, for a magnetic suspension bearing having n sets of driving windings, the phase shifting method includes: taking the carrier phase of the nth set of driving windings as a reference, the 1st, 2nd ... (n-1)th sets of windings are shifted relative to the carrier phase of the nth set of driving windings, and the phase shift angles are θ1, θ2 ... θ n-1 ;

[0041] Taking minimizing high-frequency vibration as the optimization goal, the optimal phase shift angle combination (θ1, θ2...θ n-1 ), thereby achieving high-frequency vibration suppression of the magnetic bearing.

[0042] Furthermore, the phase shift angles θ1, θ2, ... n-1 The range is [0,360).

[0043] Furthermore, the optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

[0044] The present invention also provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, comprising:

[0045] A machine learning model is constructed using a data set of PWM switching frequency and high-frequency vibration. The machine learning model is used to fit the nonlinear relationship between the PWM switching frequency and the high-frequency vibration. The trained machine learning model is used as a high-frequency vibration prediction model for obtaining a phase shift angle in a continuously changing range.

[0046] The PWM switching frequency f of the magnetic bearing is adjusted, and the adjustment method includes increasing or decreasing f;

[0047] Taking minimizing high-frequency vibration as the optimization goal, the optimal PWM switching frequency f is found through the high-frequency vibration prediction model, so as to achieve high-frequency vibration suppression of magnetic suspension bearings.

[0048] Specifically, the optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

[0049] The present invention also provides a method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, comprising:

[0050] A data set of phase shift angle, PWM switching frequency and high-frequency vibration of the carrier phase is used to construct a machine learning model, wherein the machine learning model is used to fit the nonlinear relationship between the phase shift angle, PWM switching frequency and high-frequency vibration, and the trained machine learning model is used as a high-frequency vibration prediction model for obtaining the phase shift angle and PWM switching frequency in a continuously changing range;

[0051] Carrier phase shifting is performed on each driving winding of the magnetic suspension bearing respectively; wherein, for a magnetic suspension bearing having n sets of driving windings, the phase shifting method includes: taking the carrier phase of the nth set of driving windings as a reference, the 1st, 2nd ... (n-1)th sets of windings are shifted relative to the carrier phase of the nth set of driving windings, and the phase shift angles are θ1, θ2 ... θ n-1 ; Adjust the PWM switching frequency f of the magnetic bearing, and the adjustment method includes increasing or decreasing f;

[0052] Taking minimizing high-frequency vibration as the optimization goal, the optimal phase shift angle and PWM switching frequency combination (θ1, θ2...θ n-1 , f), thereby achieving high-frequency vibration suppression of the magnetic bearing.

[0053] Specifically, the optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

[0054] The present invention also provides a high-frequency vibration suppression system for magnetic suspension bearings based on carrier phase-shift PWM, comprising: a computer-readable storage medium and a processor;

[0055] The computer-readable storage medium is used to store executable instructions;

[0056] The processor is used to read the executable instructions stored in the computer-readable storage medium to execute any of the above-mentioned methods for suppressing high-frequency vibration of magnetic suspension bearings based on carrier phase-shift PWM.

[0057] Take a typical 8-pole magnetic bearing as an example. Figure 1 As shown. A single magnetic bearing has four sets of winding PWM power supply, and the circuit used is the most common "H-bridge" topology. Taking bipolar modulation as an example, the high-frequency harmonics in the magnetic bearing winding current are mainly 1 times the switching frequency current harmonics, thus reflecting the vibration of 1 times the switching frequency (if it is unipolar modulation, it is mainly reflected as 2 times the switching frequency vibration).

[0058] The carrier phase shifting method proposed by the present invention is as follows: Figure 2 As shown, the carrier phases of the four power converters are shifted, and the carrier phase shift angle of one set of converters is used as a reference (such as the phase angle of power converter 1 is 0), then the phase difference of converters 2-4 relative to converter 1 is θ1, θ2, θ3. The change of these three phase angles from 0 to 360 degrees will not change the amplitude of the harmonic current, but will cause the spatial order of the high-frequency electromagnetic force in the air gap to change, thereby affecting the vibration amplitude of the magnetic bearing stator. For example, if there is no carrier phase shift, the carrier phase angles of the four converters are exactly the same, then the electromagnetic force dominated by "0th order" will be excited, if the carrier phase angles of the four converters are 0-0-180-180 degrees respectively, then the electromagnetic force dominated by "2nd order" will be excited, if the carrier phase angles of the four converters are 0-180-0-180 or 0-180-180-0 degrees respectively, then the electromagnetic force dominated by "1st order" will be excited. Figure 3 The figure shows the change in the amplitude of electromagnetic force of each order (v) and the excited vibration mode when the phase shift angle X changes under the 0-0-XX phase shift mode. Since the final vibration response caused by different forms of electromagnetic force is different, the final vibration size will be changed by phase shifting. Only when the phase shifting method and phase shifting angle are suitable can the vibration be minimized.

[0059] The generation of vibration originates from the process of "excitation source-transmission path-receptor". The above-mentioned change in the carrier phase shift angle is to adjust the "excitation source". In fact, the transmission path also plays an important role, which is often overlooked by researchers because the process is very complicated and difficult to model. Generally, the "mechanical transfer function" is used to describe the mapping relationship between electromagnetic force and vibration. According to the vibration superposition theorem, the vibration response is the result of the superposition of electromagnetic forces of various orders. Therefore, the mechanical transfer function is expressed in the form of "multi-peaks", and each peak represents the "resonance peak" of a certain order of electromagnetic force, such as Figure 4The resonance frequency point corresponding to the peak value depends on the natural frequency, which is related to the main structural parameters of the magnetic bearing, such as mass and stiffness. Therefore, if the external excitation frequency (the switching frequency in this example) can be adjusted to avoid the resonance point, the vibration amplitude can also be reduced.

[0060] In summary, the present invention mentions two methods for reducing the vibration magnitude, namely adjusting the phase shift angle and adjusting the switching frequency. The effect diagram is shown in FIG. Figure 4 As shown. If variable switching frequency is added on the basis of carrier phase shifting, further vibration suppression is expected to be achieved. However, this requires accurately finding the appropriate phase shifting method, the corresponding phase shifting angle, and the optimal switching frequency, which is very difficult for vibration problems. First of all, the changes in the phase shifting method and angle result in a very complex distribution of the spatial form of electromagnetic force, and the proportion of each order of electromagnetic force is difficult to determine. There is a big discrepancy between the experimental and simulation modeling, which makes it difficult to implement carrier phase shifting. Secondly, the precise form of the "mechanical transfer function" is difficult to obtain. For an actual processed magnetic bearing, its main structural parameters are difficult to obtain, resulting in the inability to accurately calculate parameters such as natural frequency and damping, making variable switching frequency difficult to implement. Therefore, the present invention proposes an artificial intelligence optimization strategy, which first predicts vibration based on machine learning, and then optimizes the PWM control parameters based on a meta-heuristic algorithm.

[0061] Specifically, we first construct a data set between the carrier phase shift angle, switching frequency and vibration magnitude through offline experiments, and then establish a deep neural network to fit the nonlinear relationship between the phase shift angle, switching frequency and vibration. Based on the generalization ability of the neural network, we obtain a vibration prediction model for the phase shift angle and switching frequency in a continuous variation range. Furthermore, based on the trained neural network model, the input of the neural network (phase shift angle, switching frequency) is optimized through the gray wolf optimization algorithm, and the optimization target is the output of the neural network (vibration), so as to find the numerical values ​​of the phase shift angle and switching frequency corresponding to the minimum vibration. The flowchart is shown in the figure below. Figure 5 shown.

[0062] To verify the effectiveness of the proposed method, experiments were carried out on a high-speed motor based on magnetic bearings. Figure 5 During the parameter optimization process, the present invention obtains the optimal parameters based on the prototype of this example: the carrier phase shift angles of the four converters are 0°, 0°, 182°, 182°, and the switching frequency is 9.66kHz. The optimal parameters are applied to the magnetic bearing controller, and the vibration comparison effect is as follows: Figure 6 shown. Figure 6(a) shows the vibration spectrum of the static suspension of the magnetic bearing rotor. It can be seen that the high-frequency vibration is concentrated at the switching frequency (9.66kHz) and its multiples. When the carrier phase-shift PWM is applied, the vibration amplitude at 9.66kHz drops from 0.935g to 0.067g, a decrease of 93%; Figure 6 (b) in the figure shows the vibration spectrum in the rotating state with a speed of 3000 rpm. Since the motor inverter is also powered by PWM and its switching frequency is 2kHz, 2kHz and its multiple frequency components appear in the spectrum. However, the 9.66kHz vibration introduced by the magnetic bearing controller is still very prominent. After applying carrier phase shift to the magnetic bearing controller, the vibration amplitude drops from 1.384g to 0.270g, a decrease of 80%, effectively improving the prominent vibration line spectrum.

[0063] At the same time, the present invention evaluates the influence of carrier shift on the operating performance of magnetic bearings, such as Figure 7 shown. Figure 7 (a) shows the change of rotor displacement in x and y directions before and after carrier phase shifting. The rotation speed is 3000 rpm. It can be seen that carrier phase shifting has no negative impact on the rotor displacement fluctuation and even improves it. Figure 7 (b) in the figure shows the effect of carrier phase shift on DC bus current. The current fluctuation amplitude is reduced from 4.85A to 0.99A, a decrease of 80%, which improves efficiency. This is mainly because the nearly 180-degree phase shift causes the high-frequency components of the bus current to superimpose and cancel each other out.

[0064] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for suppressing high-frequency vibration of a magnetic bearing based on carrier phase-shift PWM, characterized in that: include: A data set of phase shift angle and high-frequency vibration of the carrier phase is used to construct a machine learning model, wherein the machine learning model is used to fit the nonlinear relationship between the phase shift angle and the high-frequency vibration, and the trained machine learning model is used as a high-frequency vibration prediction model for obtaining the phase shift angle in a continuously changing range; Carrier phase shifting is performed on each driving winding of the magnetic suspension bearing respectively; wherein, for a magnetic suspension bearing having n sets of driving windings, the phase shifting method includes: taking the carrier phase of the nth set of driving windings as a reference, the 1st, 2nd ... (n-1)th sets of windings are shifted relative to the carrier phase of the nth set of driving windings, and the phase shift angles are θ1, θ2 ... θ n-1 ; Taking minimizing high-frequency vibration as the optimization goal, the optimal phase shift angle combination (θ1, θ2...θ n-1 ), thereby achieving high-frequency vibration suppression of the magnetic bearing.

2. The method for suppressing high-frequency vibration of a magnetic bearing based on carrier phase-shift PWM according to claim 1, characterized in that: Phase shift angle θ1, θ2...θ n-1 The range is [0,360).

3. The method for suppressing high-frequency vibration of a magnetic bearing based on carrier phase-shift PWM according to claim 1, characterized in that: The optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

4. A method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, characterized in that: include: A machine learning model is constructed using a data set of PWM switching frequency and high-frequency vibration. The machine learning model is used to fit the nonlinear relationship between the PWM switching frequency and the high-frequency vibration. The trained machine learning model is used as a high-frequency vibration prediction model for obtaining a phase shift angle in a continuously changing range. The PWM switching frequency f of the magnetic bearing is adjusted, and the adjustment method includes increasing or decreasing f; Taking minimizing high-frequency vibration as the optimization goal, the optimal PWM switching frequency f is found through the high-frequency vibration prediction model, so as to achieve high-frequency vibration suppression of magnetic suspension bearings.

5. The method for suppressing high-frequency vibration of a magnetic bearing based on carrier phase-shift PWM according to claim 4, characterized in that: The optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

6. A method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM, characterized in that: include: A data set of phase shift angle, PWM switching frequency and high-frequency vibration of the carrier phase is used to construct a machine learning model, wherein the machine learning model is used to fit the nonlinear relationship between the phase shift angle, PWM switching frequency and high-frequency vibration, and the trained machine learning model is used as a high-frequency vibration prediction model for obtaining the phase shift angle and PWM switching frequency in a continuously changing range; Carrier phase shifting is performed on each driving winding of the magnetic suspension bearing respectively; wherein, for a magnetic suspension bearing having n sets of driving windings, the phase shifting method includes: taking the carrier phase of the nth set of driving windings as a reference, the 1st, 2nd ... (n-1)th sets of windings are shifted relative to the carrier phase of the nth set of driving windings, and the phase shift angles are θ1, θ2 ... θ n-1 ; Adjust the PWM switching frequency f of the magnetic bearing, and the adjustment method includes increasing or decreasing f; Taking minimizing high-frequency vibration as the optimization goal, the optimal phase shift angle and PWM switching frequency combination (θ1, θ2...θ n-1 , f), thereby achieving high-frequency vibration suppression of the magnetic bearing.

7. The method for suppressing high-frequency vibration of a magnetic bearing based on carrier phase-shift PWM according to claim 6, characterized in that: The optimization method of the high-frequency vibration prediction model is a traversal method, a genetic algorithm, a particle swarm algorithm or a grey wolf optimization algorithm.

8. A high-frequency vibration suppression system for magnetic suspension bearings based on carrier phase-shift PWM, characterized in that: include: A computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the method for suppressing high-frequency vibration of a magnetic suspension bearing based on carrier phase-shift PWM according to any one of claims 1 to 7.

Images