An asymmetric magnetic suspension rotor bias self-decoupling vibration suppression control method

Abstract

The application discloses a kind of asymmetric magnetic suspension rotor bias self-decoupling vibration suppression control methods, it is related to magnetic suspension control technical field.The method first extracts the mechanical heterogeneity parameter of rotor system, establishes translational and rotational strong coupling benchmark;Second, break the traditional equal bias current, according to the proportion of force arm intercept inversely proportionally configure two end static bias current, realize physical self-decoupling in bottom electromagnetic actuation source;Subsequently, construct parallel double-channel composite control topology in upper differential control loop, wherein main channel is responsible for basic support, cross feedback branch is embedded in second-order high-pass filter;Finally, rely on the microsecond level parallel solution composite control current of edge side hardware platform to drive hardware actuation.The application eliminates the strong coupling caused by mechanical asymmetry from physical bottom layer, greatly reduces the computing power burden, and accurately provides lead phase compensation by virtue of double-channel topology, effectively suppresses the vortex divergence under super-speed working condition.

Description

Technical Field

[0001] This invention relates to the field of magnetic levitation dynamics control and vibration suppression technology for complex electromechanical systems, and particularly to a self-decoupling vibration suppression control method for asymmetric magnetic levitation rotor bias. Background Technology

[0002] Active magnetic bearings utilize controllable electromagnetic force to achieve contactless rotor support, offering advantages such as frictionless operation, long lifespan, and active adjustability. They are widely used in high-end energy and industrial equipment such as high-speed energy storage flywheels, centrifugal compressors, and high-power microgrid generators. Ideally, magnetically levitated rotors are often theoretically analyzed as equivalent symmetrical models; however, in practical engineering applications, due to limitations in the rotor system's mass distribution, non-uniform loads, and restricted installation space, there are often objective differences in the force arms from the magnetic bearings at both ends of the rotor to the center of mass, as well as the original displacement stiffness and current stiffness. This physical "asymmetry" breaks the independence of traditional rotor dynamics, leading to a strong cross-coupling between the rotor's radial translation and angular rotation.

[0003] Even more serious is that when such asymmetric rotors are operating at ultra-high speeds, the intense gyroscopic effect excites complex angular vortex modes, including precession modes of forward vortex and nutation modes of reverse vortex. The strong coupling caused by the asymmetry of the mechanical structure, combined with the high-frequency gyroscopic effect, not only leads to a sharp decrease in the system's damping in the nutation frequency band, making it extremely easy to induce high-frequency resonance or even instability, but also significantly increases the complexity of system dynamics modeling and full-speed stability control.

[0004] To address the strong coupling and high-frequency instability issues of asymmetric magnetically levitated rotors, existing technologies primarily focus on decoupling and compensation optimization at the control algorithm software level. Examples include employing high-dimensional multivariable robust control, state feedback based on complex observers, or online parameter identification based on real-coefficient models. However, these conventional methods generally suffer from two major limitations when implemented in practical engineering:

[0005] First, existing electromagnetic drive topologies assume that the magnetic bearings at both ends of the rotor use equal static bias currents. This homogeneous static electromagnetic actuation method cannot eliminate the torque deviation caused by the asymmetry of the mechanical lever arm at its physical root. The lack of decoupling at the underlying physical source leads to inherent limitations in the dynamic stiffness and resistance to external disturbances of the upper-level control system. Second, relying solely on upper-level algorithms for strongly coupled network compensation often introduces lengthy and complex matrix operations. Moreover, such methods are extremely sensitive to perturbations of micro-parameters and the inherent phase lag of power amplifiers, easily causing microsecond-level delays in real-time control hardware at the edge side, making it difficult to simultaneously meet the basic support stability of the low-frequency main channel and the extreme vibration suppression requirements of the high-frequency nutation mode.

[0006] Therefore, how to establish synergy between the underlying physical drive and the upper-level control topology, achieve low-cost structural decoupling of the asymmetric system from the physical source of electromagnetic actuation, and combine heterogeneous control channels to achieve efficient vibration suppression across the entire rotor speed range (especially the nutation frequency band) is an urgent technical challenge to overcome the bottleneck of high-speed stable operation of asymmetric magnetic levitation rotors and realize global system scheduling. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of the existing technologies by proposing an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method. Focusing on the challenges of strong translational-rotational coupling in asymmetric rotor structures and vortex instability induced by ultra-high-speed gyroscopes, this invention constructs a complete technical solution encompassing heterogeneous parameter extraction, static bias physical decoupling, dual-channel control setup, high-pass phase compensation, and edge hardware parallel computation. This achieves coordinated cooperation between bottom-level electromagnetic decoupling and upper-level dynamic vibration suppression, overcoming the problems of existing technologies relying solely on software algorithm compensation, being limited by inherent bias limitations, incurring high computational costs, and struggling to achieve vibration suppression across all rotational speeds.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A self-decoupling vibration suppression control method for asymmetric magnetic levitation rotor bias, characterized by comprising the following steps:

[0010] Step S1: Heterogeneous parameter perception and coupling state establishment. Perceive and extract the geometric and electromagnetic heterogeneity parameters of the asymmetric magnetic levitation rotor system in the radial degree of freedom, and establish the translational-rotational strong coupling state benchmark caused by the asymmetry of the mechanical lever arm.

[0011] Step S2: Bias current configuration and underlying physical decoupling. Based on the geometric lever arm intercepts from the upper and lower active magnetic bearings to the rotor's center of mass, configure the static bias current of the active magnetic bearings at both ends so that the ratio of the static bias current is inversely proportional to the geometric lever arm intercepts, thereby achieving electromagnetic layer self-decoupling at the physical actuation source.

[0012] Step S3: Parallel dual-channel composite control topology construction. In the upper-level differential control loop that has completed the decoupling of the underlying physical layer, a dynamically parallel heterogeneous dual-channel composite control topology is constructed. The composite control topology includes a main control channel responsible for outputting the basic control current component, and a parallel cross-feedback branch specifically for dealing with high-frequency angular eddy mode.

[0013] Step S4: High-pass cross-filtering and advanced phase compensation. A high-pass filter transfer function with high-frequency phase advance characteristics is embedded in the parallel cross-feedback branch. Using preset fixed control parameters, advanced phase compensation is provided near the high-frequency nutation mode of the rotor, and low-frequency interference signals are attenuated and isolated. The cross-feedback current component is calculated.

[0014] Step S5: Edge-side hardware calculation and full-speed vibration suppression. The rotor displacement variables are collected in real time. The calculation of the main control channel and the parallel cross-feedback branch is executed synchronously through the edge-side hardware computing platform. The calculated basic control current component and the cross-feedback current component are vector-superimposed to generate a composite differential control current. Combined with the static bias current of the underlying layer to drive the power amplifier, the eddy current divergence under high-speed rotation is suppressed.

[0015] Furthermore, the specific process of extracting geometric and electromagnetic heterogeneity parameters and establishing a strongly coupled state reference in step S1 is as follows:

[0016] Establish a coordinate system with the rotor's center of mass as the origin, and define the angular displacement variables of the rotor about the two radial axes as follows: and ;

[0017] Extract the geometric lever arm intercept from the upper active magnetic bearing to the rotor's center of mass. Geometric lever arm intercept from the lower active magnetic bearing to the rotor's center of mass and the original radial displacement stiffness of the upper and lower ends. , With radial current stiffness , ;

[0018] Constructing a local representation of the rotor angular dynamic equations that incorporates strong coupling characteristics:

[0019]

[0020] in, Let be the radial moment of inertia of the rotor. The moment of inertia is the polar rotation. This refers to the high-speed angular velocity of the rotor. and This corresponds to the control current input to the upper and lower ends.

[0021] Furthermore, the step S2, which states that the ratio of the static bias current to the geometric lever arm intercept is inversely proportional, must satisfy the following variable bias self-decoupling physical constraint equation:

[0022]

[0023] in, The static bias current injected into the upper magnetic bearing. The static bias current injected into the lower magnetic bearing;

[0024] By employing the aforementioned variable bias self-decoupling physical constraint equation, the actual electromagnetic stiffness of the upper and lower ends is adaptively biased asymmetrically, thereby satisfying... The input decoupling condition directly cancels the asymmetry of the mechanical lever arm at the electromagnetic actuation level.

[0025] Furthermore, the parallel feed-in and calculation mechanism of the heterogeneous dual-channel composite control topology in step S3 is as follows: the system obtains the real-time rotor angular displacement error signal through the displacement sensor and feeds it in parallel into the main control channel and the parallel cross-feedback branch.

[0026] Furthermore, in step S3, the main control channel is configured with a main channel controller. Used to output the basic control current component using the decoupled physical reference. It provides high dynamic foundation electromagnetic stiffness and suspension support force in the low-frequency band.

[0027] Furthermore, the parallel cross-feedback branch mentioned in step S3 is independent of the main channel and is used to calculate the output cross-feedback current component. To cope with the complex high-frequency angular vortex modes excited by the gyro effect during high-speed rotation.

[0028] Furthermore, the high-pass filter transfer function mentioned in step S4 adopts a second-order high-pass filter transfer function. Its complex field standard topological structure is constructed as follows:

[0029]

[0030] in, The high-pass cutoff angular frequency; the parallel cross-feedback branch contains a control gain connected in series. ;

[0031] The high-pass cutoff angular frequency and the control gain It can be directly called as the preset and fixed control parameter.

[0032] Furthermore, the cross-feedback branch described in step S4 utilizes the frequency domain heterogeneity of this second-order high-pass filter to provide an extremely accurate leading phase correction signal without delay near the high-frequency nutation mode frequency of the rotor; simultaneously, thanks to Due to its inherent low-frequency attenuation characteristics, the control current component generated by this branch approaches zero in the low-frequency range, automatically isolating the cross-compensation channel from interfering with the foundation support performance of the low-frequency main channel.

[0033] Furthermore, the generation and actuation process of the composite differential control current in step S5 is as follows:

[0034] The edge-side hardware computing platform, which is composed of field-programmable gate arrays and real-time controllers, collects and processes rotor displacement variables in real time with a microsecond-level control cycle.

[0035] The edge-side hardware platform uses parallel computing instructions to process the basic control current components. With the cross-feedback current component High-speed vector superposition is performed in the underlying register to generate the composite differential control current;

[0036] Furthermore, the composite differential control current mentioned in step S5 is output to a power amplifier with high-frequency phase lag characteristics via digital-to-analog conversion to drive the electromagnet coil. Through the physical superposition of the bottom static bias current and the upper dynamic composite current, the asymmetric magnetic levitation rotor can achieve stable operation across the entire speed range.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] (1) The present invention provides an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method. By introducing variable bias current self-decoupling technology, it effectively solves the multi-axis and multi-dimensional strong coupling problem of the asymmetric magnetic levitation rotor system and greatly reduces the online decoupling computing load of the control algorithm at the software level.

[0039] (2) The present invention provides an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method, which is based on a parallel heterogeneous dual-channel composite control topology dynamic collaborative support and vibration suppression circuit, ensuring that the system maintains extremely high high-frequency modal stability and dynamic stability under full speed operation conditions.

[0040] (3) The present invention provides an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method, which enhances the system’s adaptability to nutation instability caused by gyro effect at ultra-high speed and complex external disturbances by high-frequency phase accurate advance compensation and low-frequency disturbance isolation of cross channels.

[0041] (4) The present invention provides an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method, which adopts an integrated collaborative design of bottom-level physical decoupling and upper-level lightweight control, optimizes the internal signal flow, reduces control response delay, and ensures the feasibility and operating efficiency of the method in industrial edge deployment. Attached Figure Description

[0042] Figure 1 The overall flowchart of an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method provided by the present invention;

[0043] Figure 2 This is a schematic diagram of the bias self-decoupling and dual-channel composite control topology provided in an embodiment of the present invention;

[0044] Figure 3The simulation diagram shows the comparison of the dynamic decoupling effect or frequency domain characteristics of the asymmetric magnetic levitation rotor system before and after the adoption of variable bias current self-decoupling in the embodiment.

[0045] Figure 4 The simulation results show the time-domain vibration suppression comparison of the rotor's angular spatial displacement response before and after the introduction of parallel high-pass cross-feedback control. Detailed Implementation

[0046] The present invention will be further described in detail below with reference to specific embodiments.

[0047] This invention proposes a self-decoupling vibration suppression control method for asymmetric magnetic levitation rotors. The overall approach is as follows: A collaborative control architecture is designed, driven by both physical-layer electromagnetic self-decoupling and feedback-layer dual-channel composite control. This optimizes and integrates the information and control flows between physical parameter sensing, underlying electromagnetic self-decoupling, and parallel heterogeneous control at the feedback layer. This achieves real-time dynamic collaboration and integrated design between physical source decoupling and feedback loop control, thereby constructing a robust control system capable of actively eliminating structural asymmetry and precisely suppressing high-frequency eddy current modes. The ultimate goal is to fully exploit the potential of the underlying physical drive topology in eliminating strong structural coupling and improving the overall performance, operating efficiency, and high-speed stability of asymmetric magnetic levitation rotor systems, thus promoting their development towards higher levels of high-speed and stable operation.

[0048] This invention provides an asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method. By constructing a system architecture centered on variable bias current self-decoupling and parallel heterogeneous dual channels, it aims to solve the problems of limited overall control performance in current asymmetric magnetic levitation rotor systems caused by strong translational-rotational coupling due to mechanical structural asymmetry, and high-frequency nutation instability caused by gyroscopic effects at ultra-high speeds. Specifically, this method utilizes a heterogeneous configuration where the ratio of the front-end and rear-end static bias currents is inversely proportional to the lever arm intercept, thus canceling the mechanical lever arm asymmetry at the electromagnetic force generation level, transforming the system into a structurally decoupled dynamic model at the underlying level. Through the coordinated operation of the main channel controller and the parallel branch cross-feedback compensation channels, it provides precise advance phase compensation near the rotor nutation frequency and attenuates low-frequency disturbances, improving the system's hard stability margin and adaptive capability to uncertainties such as modal divergence and complex external disturbances under ultra-high-speed operating conditions. Furthermore, it achieves real-time parallel calculation and actuation control of composite control currents within the system, reducing control decision delay, thereby enhancing the feasibility and operational efficiency of this method in real-time control hardware environments at the industrial edge.

[0049] The specific technical solution is as follows, including the following detailed steps:

[0050] Step 1: Heterogeneous parameter sensing and coupling state establishment

[0051] The geometric and electromagnetic heterogeneity parameters of the asymmetric magnetic levitation rotor system in the radial degree of freedom are sensed and extracted. Specifically, a coordinate system with the rotor's center of mass as the origin is established, and the angular displacement variables of the rotor about the two radial axes are defined as follows: and Extract the geometric lever arm intercept from the upper active magnetic bearing to the rotor's center of mass. Geometric lever arm intercept from the lower active magnetic bearing to the rotor's center of mass Simultaneously, the basic electromagnetic parameters of the system under nominal conditions are extracted.

[0052] For mechanical asymmetry (i.e.) The translational and rotational disturbances caused by the asymmetry of the lever arm are used to establish a reference state for the strong translational-rotational coupling due to lever arm asymmetry. Under the traditional homogeneous bias current-driven topology, its rotor angular dynamic equation can be locally characterized as:

[0053]

[0054] in, Let be the radial moment of inertia of the rotor. The moment of inertia is the polar rotation. This refers to the high-speed angular velocity of the rotor. and These are the original radial displacement stiffnesses at the upper and lower ends, respectively. and These are the radial current stiffnesses at the upper and lower ends, respectively. and To correspond to the control current input to the upper and lower ends, a reference state for the strong translational-rotational coupling caused by lever arm asymmetry is established.

[0055] Step 2: Bias current configuration and underlying physical decoupling

[0056] This design breaks away from the traditional homogeneous topology configuration that uses the same static bias current at both ends. To offset the difference in lever arm intercept from the physical actuation source, the static bias current of the active magnetic bearings at both ends is adaptively configured so that its static current ratio is strictly inversely proportional to the mechanical lever arm intercept, thus satisfying the variable bias self-decoupling physical constraint equation:

[0057]

[0058] in, The static bias current injected into the upper magnetic bearing. The static bias current injected into the lower magnetic bearing.

[0059] By configuring the variable bias current as described above, an adaptive asymmetric bias can be generated between the actual electromagnetic stiffness of the upper and lower ends, thereby satisfying the requirements. The input decoupling conditions.

[0060] This physical reverse compensation mechanism actively bears the cost of static instability distortion derived from asymmetric displacement stiffness, so that the translational resultant force and rotational control torque generated by the rotor under the action of controlled differential current are completely decoupled and isolated in space, directly eliminating the mechanical lever arm asymmetry at the electromagnetic actuation level.

[0061] Step 3: Construction of Parallel Dual-Channel Composite Control Topology

[0062] Based on the decoupling of the underlying electromagnetic source completed in step 2, a dynamically parallel heterogeneous dual-channel composite control topology is constructed in the upper-level differential control loop of the system. After acquiring the real-time rotor angular displacement error signal through a displacement sensor, the system feeds it in parallel into the two heterogeneous channels:

[0063] (1) Main control channel: Configure the main channel controller (Such as a PID controller with an incomplete differential network), the basic control current component is specifically calculated and output using the decoupled physical reference. It provides high dynamic electromagnetic stiffness and stable suspension support force in the low-frequency band.

[0064] (2) Parallel cross-feedback branch: A high-frequency compensation loop is constructed independently of the main channel to calculate the output cross-feedback current component. Specifically designed to handle high-speed rotation. Complex high-frequency angular vortex modes excited by the term (i.e., the gyro effect term).

[0065] Step 4: High-pass cross-filtering and lead phase compensation

[0066] In the parallel cross-feedback branches, a second-order high-pass filter transfer function with high-frequency phase lead characteristics is explicitly embedded. Its complex field standard topological structure is constructed as follows:

[0067]

[0068] in, This is the high-pass cutoff angular frequency. The control gain of the cross channel is connected in series, and both are directly called as preset fixed parameters.

[0069] The solidification parameters can be pre-configured based on the rigid phase margin constraints required by the system, using offline analytical algorithms such as complex plane coincidence or geometric mapping.

[0070] Utilizing the frequency domain heterogeneity of this second-order high-pass filter, the cross-feedback branch can provide an extremely accurate leading phase correction signal without delay near the rotor's high-frequency nutation mode (reverse eddy) frequency. Simultaneously, thanks to... Due to its inherent low-frequency attenuation characteristics, the control current component generated by this branch approaches zero in the low-frequency range, thereby automatically isolating the cross-compensation channel from the basic support performance of the low-frequency main channel at the signal level.

[0071] Step 5: Edge-side hardware calculation and full-speed vibration suppression

[0072] Relying on an edge-side hardware computing platform consisting of a field-programmable gate array (FPGA) and a real-time (RT) controller, the control system acquires and processes rotor displacement variables in real time with microsecond-level control cycles. Thanks to the fact that the underlying bias current in step 2 has absorbed the computational burden of strong coupling, the system's online computing power is highly concentrated.

[0073] The edge hardware platform uses parallel computing instructions to synchronously execute the calculations of the main control channel and the cross-feedback branch, and then processes the calculated basic control current components. With cross-feedback current components High-speed vector superposition is performed in the underlying register to generate the final composite differential control current.

[0074] The current is output to a power amplifier with high-frequency phase lag characteristics via digital-to-analog conversion. Combined with the static bias current of the underlying asymmetric configuration, it drives the electromagnet coil, completely suppressing eddy current divergence under high-speed rotation, and realizing the rotor's extreme vibration suppression and stable operation across the entire speed range.

[0075] Combination Figure 2 The block diagram of the asymmetric magnetic levitation rotor bias self-decoupling and dual-channel composite control system shown further elaborates on the system control flow and actuation logic of the present invention.

[0076] During system operation, the real-time spatial angular displacement feedback signal of the rotor is first obtained through the displacement sensor. and compare it with the set target angular displacement reference value. Perform a difference comparison to generate an angular displacement error signal. This error signal serves as the sole dynamic input excitation for the subsequent composite control loop.

[0077] Acquire error signal Then, the system feeds the signal in parallel into two heterogeneous computing branches:

[0078] On the one hand, the error signal is input to the main control channel and then processed by the main channel controller. A proportional-integral-differential (PI-DI) computational network is used to calculate and output the basic control current components. This component is primarily responsible for providing stiffness and foundation suspension support in the low-frequency range.

[0079] On the other hand, the error signal is synchronously fed into the cross-feedback branch in parallel. This branch contains a second-order high-pass filter with high-frequency phase lead characteristics. Targeted filtering and gain amplification are performed on the rotor angular whirl mode, and the cross-feedback current component of the output is calculated. .

[0080] To address the extremely narrow nutation divergence time window under ultra-high-speed operating conditions, in practical implementation, the edge-side hardware computing platform can rely on an industrial-grade real-time simulation and control system to support the aforementioned logic. Through microsecond-level control cycles, the basic control current component output from the main channel and the cross-feedback current component output from the cross-feedback branch are rapidly vector-summed in the underlying registers to generate a dynamic composite differential control current. .

[0081] Finally, this dynamic composite current is delivered to the power amplifier. Simultaneously, according to the variable bias self-decoupling physical constraint equation... Pre-calculated bottom static bias current It is also injected into the power amplifier. The power amplifier physically superimposes the static bias and the dynamic composite current, outputting the final drive current to the electromagnet coil of the asymmetric magnetic levitation rotor system. This process cancels the off-center load generated by the rotor's asymmetric lever arm at the electromagnetic actuation source, and, combined with the advanced phase compensation of the parallel channel, completely suppresses the high-frequency eddy current divergence induced by the gyro effect.

[0082] Example:

[0083] This embodiment provides a real-time simulation scheme for an extremely asymmetric magnetic levitation rotor under ultra-high-speed operating conditions to verify the effectiveness of the above-mentioned bias self-decoupling vibration suppression control method.

[0084] The core parameters of the simulation asymmetric magnetic levitation rotor model are set as follows:

[0085] Distance from the upper magnetic bearing to the rotor's center of mass The distance from the lower magnetic bearing to the rotor's center of mass As can be seen from the above parameters, the difference in lever arms between the upper and lower ends of the system is as high as 33.8 times, which belongs to an extremely asymmetric physical structure that is very prone to strong translational-rotational coupling.

[0086] First, the underlying variable bias self-decoupling scheme is verified.

[0087] In traditional control topologies, the upper and lower magnetic bearings are typically configured with equal static bias currents. (See attached...) Figure 3 As shown by the dashed line, due to the huge difference in physical lever arms, the cross-coupling transfer function of the system's translational displacement versus angular motion has a very large amplitude, and the strong coupling characteristics are significant.

[0088] By employing the variable bias current electromagnetic layer self-decoupling method described in step 2 of this invention, the bias currents at both ends are strictly configured according to the inverse lever arm ratio, thus satisfying... As attached Figure 3 As shown by the solid line in the figure, after introducing physical layer reverse compensation, the magnitude of the original translational-rotational strong coupling term is greatly reduced, and the asymmetry of the mechanical structure is directly canceled at the electromagnetic actuation level. The system is successfully equivalent to a decoupled object, which greatly reduces the computational load of subsequent online decoupling.

[0089] Secondly, the ultra-high-speed vibration suppression effect of the dual-channel composite control topology was verified.

[0090] The rotor is set to face an extremely demanding 24,000 rpm (i.e.) In ultra-high-speed operation, the nutation mode excited by the gyro effect can easily lead to system instability. In this embodiment, the control algorithm is deployed on an edge-side hardware real-time simulation platform that includes a field-programmable gate array (FPGA) and a real-time (RT) controller, and the control instructions are executed in parallel at microsecond intervals.

[0091] As attached Figure 4 As shown in the time-domain vibration comparison waveforms, without the introduction of a cross-feedback branch (relying only on the main channel differential control), the rotor's angular spatial displacement rapidly diverges after encountering a disturbance due to high-frequency phase lag (dashed line), posing a risk of shaft failure. However, after introducing the heterogeneous dual-channel composite control topology constructed in this invention (solid line), thanks to the computing power released by the underlying decoupling and the precise leading phase compensation current injected by the cross channel in the nutation frequency band, the rotor's angular amplitude is rapidly and effectively suppressed within tens of milliseconds and smoothly converges to a steady state. This strongly demonstrates that even under extremely harsh ultra-high-speed operating conditions, this invention can still achieve microsecond-level real-time ultra-fast operation and efficient vibration suppression through hardware and software collaboration and a lightweight architecture.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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 self-decoupling vibration suppression control method for asymmetric magnetic levitation rotor bias, characterized in that, Includes the following steps: Step S1: Heterogeneous parameter perception and coupling state establishment. Perceive and extract the geometric and electromagnetic heterogeneity parameters of the asymmetric magnetic levitation rotor system in the radial degree of freedom, and establish the reference for the strong coupling state of translation and rotation caused by the asymmetry of the mechanical lever arm. Step S2: Bias current configuration and underlying physical decoupling. Based on the geometric lever arm intercepts from the upper and lower active magnetic bearings to the rotor's center of mass, configure the static bias current of the active magnetic bearings at both ends so that the ratio of the static bias current is inversely proportional to the geometric lever arm intercepts, thereby achieving electromagnetic layer self-decoupling at the physical actuation source. Step S3: Parallel dual-channel composite control topology construction. In the upper-level differential control loop that has completed the decoupling of the underlying physical layer, a dynamically parallel heterogeneous dual-channel composite control topology is constructed. The composite control topology includes a main control channel responsible for outputting the basic control current component, and a parallel cross-feedback branch specifically for dealing with high-frequency angular eddy mode. Step S4: High-pass cross-filtering and advanced phase compensation. A high-pass filter transfer function with high-frequency phase advance characteristics is embedded in the parallel cross-feedback branch. Using preset fixed control parameters, advanced phase compensation is provided near the high-frequency nutation mode of the rotor, and low-frequency interference signals are attenuated and isolated. The cross-feedback current component is calculated. Step S5: Edge-side hardware calculation and full-speed vibration suppression. The rotor displacement variables are collected in real time. The calculation of the main control channel and the parallel cross-feedback branch is executed synchronously through the edge-side hardware computing platform. The calculated basic control current component and the cross-feedback current component are vector-superimposed to generate a composite differential control current. Combined with the static bias current of the underlying layer to drive the power amplifier, the eddy current divergence under high-speed rotation is suppressed.

2. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The specific process of extracting geometric and electromagnetic heterogeneity parameters and establishing a strongly coupled state reference in step S1 is as follows: Establish a coordinate system with the rotor's center of mass as the origin, and define the angular displacement variables of the rotor about the two radial axes as follows: and ; Extract the geometric lever arm intercept from the upper active magnetic bearing to the rotor's center of mass. Geometric lever arm intercept from the lower active magnetic bearing to the rotor's center of mass and the original radial displacement stiffness of the upper and lower ends. , With radial current stiffness , ; Constructing a local representation of the rotor angular dynamic equations that incorporates strong coupling characteristics: in, Let be the radial moment of inertia of the rotor. The moment of inertia is the polar rotation. This refers to the high-speed angular velocity of the rotor. and This corresponds to the control current input to the upper and lower ends.

3. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The step S2, which states that the ratio of the static bias current to the geometric lever arm intercept must satisfy the following variable bias self-decoupling physical constraint equation: in, The static bias current injected into the upper magnetic bearing. The static bias current injected into the lower magnetic bearing; The variable bias self-decoupling physical constraint equation enables an adaptive asymmetric bias in the actual electromagnetic stiffness of the upper and lower ends, satisfying the following conditions: The input decoupling condition directly cancels the asymmetry of the mechanical lever arm at the electromagnetic actuation level.

4. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The parallel feed and calculation mechanism of the heterogeneous dual-channel composite control topology in step S3 is as follows: the system obtains the real-time rotor angular displacement error signal through the displacement sensor and feeds it in parallel into the main control channel and the parallel cross feedback branch.

5. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The main control channel configuration in step S3 is a main channel controller, which is used to output the basic control current component using the decoupled physical reference, providing high dynamic basic electromagnetic stiffness and suspension support force in the low frequency band.

6. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The parallel cross-feedback branch mentioned in step S3 is independent of the main channel and is used to calculate the output cross-feedback current component to deal with the complex high-frequency angular eddy mode excited by the gyro effect during high-speed rotation.

7. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The high-pass filter transfer function described in step S4 adopts a second-order high-pass filter transfer function, and its standard topological structure in the complex domain is constructed as follows: in, The high-pass cutoff angular frequency; the parallel cross-feedback branch contains a control gain connected in series. ; The high-pass cutoff angular frequency and the control gain The second-order high-pass filter transfer function is directly invoked as the preset, fixed control parameter. As an abbreviation.

8. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The cross-feedback branch in step S4 utilizes the frequency domain heterogeneity of this second-order high-pass filter to provide a leading phase correction signal without delay near the high-frequency nutation mode frequency of the rotor; simultaneously, thanks to Due to its inherent low-frequency attenuation characteristics, the control current component generated by this branch approaches zero in the low-frequency range, automatically isolating the cross-compensation channel from interfering with the foundation support performance of the low-frequency main channel.

9. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, The generation and actuation process of the composite differential control current in step S5 is as follows: The edge-side hardware computing platform, which is composed of field-programmable gate arrays and real-time controllers, collects and processes rotor displacement variables in real time with a microsecond-level control cycle. The edge hardware platform uses parallel computing instructions to perform high-speed vector superposition of the basic control current component and the cross-feedback current component in the underlying register to generate the composite differential control current.

10. The asymmetric magnetic levitation rotor bias self-decoupling vibration suppression control method according to claim 1, characterized in that, In step S5, the composite differential control current is output to a power amplifier with high-frequency phase lag characteristics via digital-to-analog conversion to drive the electromagnet coil. Through the physical superposition of the bottom static bias current and the upper dynamic composite current, the asymmetric magnetic levitation rotor can achieve stable operation across the entire speed range.

Images