A method and system for multi-axis coordinated control of a magnetic levitation motor

Abstract

The application discloses a kind of magnetic suspension motor multi-axis coordination control method and system, by gathering multi-axis current feedback signal and radial displacement signal, current signal is carried out PWM modulation processing, position deviation parameter is extracted from displacement signal, and multi-degree-of-freedom control response value is generated by coupling association;Decoupling control sequence is constructed and stability is detected, and coupling coefficient sample set is extracted by interaxle coupling analysis, and control parameter coordination matrix is constructed;Control efficiency characteristic analysis is identified high-efficiency control area, and work point position is determined in combination with load characteristic identification, and coordination control node is determined by comparison;Conversion parameter is corrected and determined by control weight distribution, and multi-axis coordination sequence is formed using power switch controller, and PWM signal synchronization generates coordination control strategy;Dynamic characteristic data is obtained by response speed analysis on current signal, control response period is determined, synchronization conversion operation is executed to generate coordination control instruction, and motor multi-axis coordination control is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of magnetic suspension control, in particular to a magnetic suspension motor multi-axis coordinated control method and system. BACKGROUND

[0002] As the core component of high-precision control equipment, magnetic suspension motors play a key role in precision machining, high-speed rotating machinery, aerospace, etc. Magnetic suspension motors achieve the non-contact suspension of the rotor through electromagnetic force, which requires precise coordinated control of multi-axis electromagnets to maintain the stable suspension state of the rotor. However, the existing magnetic suspension motor control method has significant deficiencies in multi-axis coordinated control.

[0003] Traditional magnetic suspension control methods usually adopt independent control of each axis, ignoring the coupling relationship and coordination requirements between axes. This control method leads to a lack of effective information exchange and coordination mechanism between axis controllers, and when the system load changes or external disturbances occur, the axis controllers cannot respond effectively, resulting in decreased control accuracy, poor system stability, high energy consumption, etc. Especially in complex working conditions, the coupling effect between axes will further worsen the control effect, and even lead to system instability. The existing technology lacks deep analysis and coordinated processing capability of multi-axis current signals, and cannot fully tap the potential of coordinated control between axes, making it difficult to achieve system-level optimized control effect. SUMMARY

[0004] The present application discloses a magnetic suspension motor multi-axis coordinated control method and system, which realizes the coupling and correlation processing of PWM modulated current signals and position deviation parameters through comprehensive analysis of multi-axis current feedback signals and radial displacement signals, generates multi-degree-of-freedom control response values, and then constructs decoupling control sequences and coordinated control nodes, determines conversion parameters and forms multi-axis coordinated sequences, finally generates accurate coordinated control instructions through dynamic characteristic analysis and synchronous conversion operation, realizes high-precision coordinated control between axes of the magnetic suspension motor, and improves the control performance, stability and energy efficiency of the system.

[0005] The present application discloses a magnetic suspension motor multi-axis coordinated control method and system, which realizes the coupling and correlation processing of PWM modulated current signals and position deviation parameters through comprehensive analysis of multi-axis current feedback signals and radial displacement signals, generates multi-degree-of-freedom control response values, and then constructs decoupling control sequences and coordinated control nodes, determines conversion parameters and forms multi-axis coordinated sequences, finally generates accurate coordinated control instructions through dynamic characteristic analysis and synchronous conversion operation, realizes high-precision coordinated control between axes of the magnetic suspension motor, and improves the control performance, stability and energy efficiency of the system.

[0006] The present application discloses a magnetic suspension motor multi-axis coordinated control method and system, which realizes the coupling and correlation processing of PWM modulated current signals and position deviation parameters through comprehensive analysis of multi-axis current feedback signals and radial displacement signals, generates multi-degree-of-freedom control response values, and then constructs decoupling control sequences and coordinated control nodes, determines conversion parameters and forms multi-axis coordinated sequences, finally generates accurate coordinated control instructions through dynamic characteristic analysis and synchronous conversion operation, realizes high-precision coordinated control between axes of the magnetic suspension motor, and improves the control performance, stability and energy efficiency of the system.

[0007] construct a decoupling control sequence by using the multi-degree-of-freedom control response value, perform stability detection on the decoupling control sequence to obtain different axial data, perform inter-axis coupling analysis by using the different axial data to extract a coupling coefficient sample set, and construct a control parameter coordination matrix based on the coupling coefficient sample set;

[0008] perform control efficiency feature analysis on the control parameter coordination matrix to identify a high-efficiency control area, determine a current load mode based on load feature identification of the multi-axis current feedback signal, select a mode matching working point position from the high-efficiency control area according to the current load mode, and determine a coordination control node by comparing the mode matching working point position with the multi-degree-of-freedom control response value;

[0009] determine a conversion parameter set by control weight distribution correction according to the coordination control node, use the conversion parameter to form a multi-axis coordination sequence through a power switch controller, perform multi-axis PWM signal synchronization on the multi-axis coordination sequence to generate a multi-axis coordination control strategy;

[0010] perform response speed analysis on the multi-axis current feedback signal to obtain dynamic characteristic data, determine a control response period based on the dynamic characteristic data and the multi-axis coordination control strategy, and generate a coordination control instruction by performing a multi-axis synchronous conversion operation according to the control response period.

[0011] The second aspect of the present application proposes a multi-axis coordination control system of a magnetic suspension motor, comprising:

[0012] A signal acquisition module is configured to acquire multi-axis current feedback signals and radial displacement signals of the magnetic suspension motor, perform multi-channel PWM modulation on the multi-axis current feedback signals to obtain PWM modulated current signals, extract position deviation parameters from the radial displacement signals, and generate multi-degree-of-freedom control response values by coupling and correlating the PWM modulated current signals and the position deviation parameters.

[0013] A coupling analysis module is configured to construct a decoupling control sequence by using the multi-degree-of-freedom control response value, perform stability detection on the decoupling control sequence to obtain different axial data, perform inter-axis coupling analysis by using the different axial data to extract a coupling coefficient sample set, and construct a control parameter coordination matrix based on the coupling coefficient sample set.

[0014] A load identification module is configured to perform control efficiency feature analysis on the control parameter coordination matrix to identify a high-efficiency control area, determine a current load mode based on load feature identification of the multi-axis current feedback signal, select a mode matching working point position from the high-efficiency control area according to the current load mode, and determine a coordination control node by comparing the mode matching working point position with the multi-degree-of-freedom control response value.

[0015] The control coordination module is configured to determine a conversion parameter set according to a control weight distribution correction of the coordination control node, generate a multi-axis coordination sequence by using the power switch controller according to the conversion parameter, generate a multi-axis coordination control strategy by synchronously generating multi-axis PWM signals according to the multi-axis coordination sequence, and perform multi-axis synchronous conversion operation according to dynamic characteristic data obtained by responding speed analysis of the multi-axis current feedback signals to determine a control response period and generate a coordination control instruction.

[0016] The dynamic adjustment module is configured to obtain dynamic characteristic data by responding speed analysis of the multi-axis current feedback signals, determine a control response period based on the dynamic characteristic data and the multi-axis coordination control strategy, and perform multi-axis synchronous conversion operation to generate a coordination control instruction.

[0017] The beneficial effects of the present application are embodied in the following aspects: first, by synchronously collecting and processing the multi-axis current feedback signals and the radial displacement signals, the PWM modulated current signals are coupled and associated with the position deviation parameters to generate multi-degree-of-freedom control response values, and then a decoupling control sequence is constructed and a control parameter coordination matrix is established by inter-axis coupling analysis, thereby realizing the organic combination of electromagnetic control information and position feedback information, converting the complex multi-axis coupling control problem into control units that are relatively independent and maintain a coordinated relationship, simplifying the design complexity of the control algorithm, ensuring the coordinated control effect between axes, and improving the control accuracy and operation stability of the magnetic levitation system. Second, the control efficiency characteristic analysis method is used to identify the high-efficiency control region, the current load mode is determined by combining the load characteristic identification technology, the optimal working point position is selected by mode matching, and the coordination control node is determined by comparing with the multi-degree-of-freedom control response values. This adaptive control strategy selection mechanism can dynamically adjust the control parameters according to the actual working state and load conditions of the system, select the corresponding optimal control strategy under different load modes, and realize the balanced optimization of control performance and energy efficiency. Finally, based on dynamic response analysis and synchronous control technology, the conversion parameter is determined by control weight distribution correction, the multi-axis coordination sequence is generated by using the power switch controller, the synchronous control of the multi-axis PWM signals is realized, the dynamic characteristic data is obtained by responding speed analysis of the multi-axis current feedback signals, the control response period is determined and the synchronous conversion operation is performed to generate the coordination control instruction, which can accurately coordinate the time sequence according to the actual response characteristics of each axis, ensuring the synchronous execution and coordinated cooperation of the multi-axis control instructions.

[0018] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0019] The drawings herein show specific examples of the technical solutions described in the present application, and constitute part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0020] Figure 1It is a flow diagram of a kind of magnetic suspension motor multi-axis coordinated control method of the application.

[0021] Figure 2 It is the structure block diagram of a kind of magnetic suspension motor multi-axis coordinated control system of the application. DETAILED DESCRIPTION

[0022] In the following description, for the sake of explanation and not limitation, specific details are set forth, such as specific system structures, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can be practiced in other embodiments that do not conform to these specific details. In other instances, well-known structures, devices, circuits, and methods have not been described in detail in order to avoid unnecessarily obscuring the description of the present application.

[0023] It should be understood that the term "comprises" when used in this specification and the appended claims indicates the presence of the described features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0024] Reference in the specification to "one embodiment" or "some embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. The appearances of the phrase "in one embodiment" or "in some embodiments" in various places in the specification are not necessarily all referring to the same embodiment, although it can. The terms "including," "comprising," "having," and variations thereof are meant to encompass the presence of the stated feature, step, operation, element, and / or component, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0025] The technical solutions of the embodiments of the present application are introduced below.

[0026] As shown in Figure 1 The embodiment of the application provides a kind of magnetic suspension motor multi-axis coordinated control method, comprising the following steps S110-S150:

[0027] Step S110, the multi-axis current feedback signal and radial displacement signal of magnetic suspension motor are collected, the multi-channel PWM modulation current signal is obtained by carrying out multi-channel PWM modulation to multi-axis current feedback signal, position deviation parameter is extracted from radial displacement signal, and the multi-degree-of-freedom control response value is generated by coupling correlation of PWM modulation current signal and position deviation parameter.

[0028] Specifically, multi-axis current feedback signals and radial displacement signals of the magnetic suspension motor are collected. Hall current sensors are installed in the X-axis, Y-axis and Z-axis excitation winding loops of the magnetic suspension motor to monitor the working current changes of the three-axis electromagnets in real time. For example, in the suspension control system of a high-speed magnetic suspension train, multiple groups of electromagnets are arranged below each carriage, and the excitation currents of the electromagnets need to be accurately monitored to maintain stable suspension. The sensor uses a closed-loop Hall device, which has good linearity and temperature stability, and can maintain measurement accuracy in complex electromagnetic working conditions. The current signal is transmitted to the signal conditioning module through a shielded cable, and after differential amplification and anti-aliasing filtering, it is converted into a standard analog voltage signal. The data acquisition system uses high-speed synchronous sampling to ensure the capture of transient characteristics and high-frequency components of the current waveform. In the application of magnetic suspension bearings, the electromagnets need to respond quickly to position changes when the rotor rotates at high speed, and the collection of current signals must have sufficient bandwidth and accuracy. Synchronous collection of radial displacement signals uses four eddy current displacement sensors installed in a quadrature configuration around the rotor to form a complete radial displacement monitoring system. The sensor works in a high-frequency excitation mode to detect the eddy current effect on the metal rotor surface and outputs a linear signal proportional to the displacement. The rotor radial displacement signal shows micron-level position fluctuations, reflecting the dynamic balance state of the magnetic suspension system and the influence of external disturbances.

[0029] The multi-channel PWM modulation is performed on the multi-axis current feedback signals to obtain PWM modulated current signals. The collected X-axis, Y-axis and Z-axis current feedback signals are sent into corresponding PWM modulators for independent pulse width modulation processing. Each PWM modulator generates a corresponding pulse width modulation waveform according to the amplitude and frequency characteristics of the current feedback signal. For example, in a magnetic suspension bearing system, three-axis electromagnets require independent PWM drive signals, and the duty cycle and frequency of each signal need to be accurately controlled. The PWM modulation uses a triangular wave carrier method, and the carrier frequency is set in the ultrasonic frequency band to reduce audio noise. After receiving the current feedback signal, the X-axis PWM modulator compares the signal with the triangular carrier through a comparator to generate a corresponding pulse width modulation signal. The phase synchronization between channels is maintained during the modulation process to avoid electromagnetic interference caused by phase differences. The Y-axis and Z-axis PWM modulators use the same modulation strategy, but the carrier phases lag by 120 degrees and 240 degrees, respectively, forming a three-phase symmetric modulation mode. In the suspension electromagnet control of a magnetic suspension train, multi-channel PWM modulation ensures the independent control and coordinated work of each electromagnet. The PWM modulated current signal output by the modulator has a constant switching frequency and a variable pulse width, and the pulse width directly corresponds to the excitation current size of the electromagnet. The PWM signal is driven by a power amplifier to drive the electromagnet winding, realizing accurate control of electromagnetic force.

[0030] The position deviation parameter is extracted from the radial displacement signal. The four-path radial displacement sensor signals collected are processed to calculate the deviation of the actual position of the rotor from the set center position. The measurement values of the four sensors are converted into X and Y direction deviations in the Cartesian coordinate system through coordinate transformation. For example, in a magnetic levitation flywheel energy storage system, the radial displacement of the rotor directly affects the dynamic balance and running stability of the system, and the position deviation information needs to be accurately extracted. The extraction of the position deviation parameter uses a weighted average algorithm, which assigns different weight coefficients according to the measurement accuracy and reliability of each sensor. The measurement values of the 0-degree and 180-degree sensors are used to calculate the X-direction position deviation, and the measurement values of the 90-degree and 270-degree sensors are used to calculate the Y-direction position deviation. The installation error and zero drift of the sensors are considered in the calculation process, and are compensated and corrected by calibration coefficients. The extracted position deviation parameters include three key information: radial deviation amplitude, deviation direction angle, and deviation change rate. In a magnetic levitation precision machine tool spindle, even a micron-level position deviation can affect the machining precision, so the extraction accuracy of the deviation parameter is extremely high.

[0031] In some embodiments, the coupling of the PWM modulated current signal and the position deviation parameter generates a multi-degree-of-freedom control response value, including: performing magnetic field saturation state detection on the PWM modulated current signal to generate a saturation state signal and a linear state signal; using the linear state signal to perform nonlinear correction processing on the saturation state signal to form a corrected current signal; using the position deviation parameter to perform shaft control weight distribution on the corrected current signal to generate a distributed current vector; and based on the distributed current vector, implementing magnetic field saturation compensation control to obtain a multi-degree-of-freedom control response value.

[0032] For example, the magnetic field saturation state detection on the PWM modulated current signal to generate a saturation state signal and a linear state signal includes: obtaining current distortion characteristics affected by the switching dead zone based on the PWM modulated current signal; determining actual dead zone time parameters by dead zone time back calculation on the current distortion characteristics; generating a dead zone compensation signal according to the actual dead zone time parameters; and generating a saturation state signal and a linear state signal by control compensation superposition of the dead zone compensation signal and the PWM modulated current signal.

[0033] The current distortion characteristics affected by the switching dead time are obtained based on the PWM modulated current signal. The waveform of the PWM modulated current signal is analyzed, and the current discontinuity characteristics at the switching commutation moment are identified. The complete waveform of the PWM modulated current signal is collected through high-speed sampling, and the time point at which the current slope suddenly changes is located in each switching cycle. The amplitude difference ΔI before and after the sudden change is extracted as the distortion amplitude characteristic. The variation law of the current waveform near the switching commutation point is analyzed, and the duration Δt of the abnormal change of the current slope is measured as the distortion time characteristic. For example, in the PWM driver of a high-power magnetic levitation system, by comparing the ideal current waveform with the actual PWM modulated current signal, three key distortion characteristic parameters, including the current step height, step width, and slope change rate, are quantitatively extracted. The frequency domain analysis method is used to separate the characteristic harmonic components of the dead-time distortion from the PWM modulated current signal, and the characteristic spectral lines at the odd multiple frequencies of the switching frequency are identified. In the high-frequency PWM modulation system, the correspondence between the distortion characteristic parameters and the PWM signal parameters is established, and a distortion characteristic database is formed, which contains the current distortion amplitude, time characteristic, and spectral characteristic under different working conditions. The extracted distortion characteristic parameters are arranged as structured data.

[0034] The actual dead-time parameters are determined by the dead-time time reverse calculation based on the current distortion characteristics. The current distortion amplitude ΔI and the distortion time characteristic Δt extracted in the previous section are used to establish a dead-time calculation model. The relationship between the dead-time and the distortion characteristics is established based on circuit theory: Tdead=(ΔI×L) / Vdc, where ΔI is the current step amplitude extracted in the previous section, L is the load inductance, and Vdc is the DC bus voltage. The distortion amplitude ΔI measured in the previous section is substituted into the formula to calculate the preliminary dead-time estimate. Cross-validation is performed in combination with the distortion time characteristic Δt extracted in the previous section, and the calculation accuracy is evaluated by comparing the calculation results with the measurement results. Considering the influence of the discreteness and temperature drift of the device characteristic parameters on the dead-time, a correction coefficient is introduced based on the spectral characteristic information extracted in the previous section to improve the calculation accuracy. The distortion characteristic database established in the previous section is used for table lookup correction, and the closest reference data is matched according to the distortion characteristics of the current working condition. The asymmetry of the dead-time of the upper and lower bridge arm switching devices is analyzed, and the corresponding dead-time parameters are calculated using the distortion characteristics of the forward and reverse currents, respectively. The actual dead-time parameters calculated based on the distortion characteristics are finally confirmed to form accurate parameter values.

[0035] The dead-time compensation signal is generated based on the actual dead-time parameter. Based on the actual dead-time parameter Tdead calculated in the previous step, a compensation signal generator is designed to generate a compensation pulse with a duration of Tdead. The amplitude of the compensation pulse is calculated based on the actual dead-time parameter and circuit parameters: Icomp = (Vdc x Tdead) / L, where Icomp is the compensation current amplitude, Vdc is the DC bus voltage, and L is the load inductance. For example, in a precision magnetic levitation positioning system, when the actual dead-time parameter is 2.5 microseconds, the corresponding compensation current amplitude needs to be accurately matched to this time parameter to eliminate current distortion. According to the value of the actual dead-time parameter, the pulse width and trigger time of the compensation signal are dynamically adjusted to ensure that the compensation effect accurately matches the actual dead-time characteristics. Considering the dynamic characteristics of the circuit system, the actual dead-time parameter is substituted into the waveform shaping algorithm to make the compensation pulse match the natural response characteristics of the system. In high-speed switching PWM systems, the rise and fall times of the compensation pulse are optimized based on the actual dead-time parameter to avoid generating new oscillations. By using the actual dead-time parameter as the core parameter, real-time generation and adaptive adjustment of the compensation signal are achieved, supporting dynamic optimization based on changes in the dead-time. The frequency spectrum characteristics of the compensation signal are designed based on the actual dead-time parameter, and the best compensation effect is achieved through the principle of destructive interference in the frequency domain.

[0036] The saturation state signal and linear state signal are generated by superimposing the dead-time compensation signal and the PWM modulated current signal. The generated dead-time compensation signal is accurately superimposed in the time domain with the original PWM modulated current signal. The superposition operation is implemented using digital signal processing technology to ensure perfect alignment of the two signals in time and phase. For example, in a magnetic levitation flywheel system, the current demand of the rotor at different speeds varies greatly, and the effect of dead-time compensation also changes accordingly, requiring real-time monitoring of the superposition effect. The amplitude and frequency spectrum of the signal are monitored during the superposition process to observe the actual improvement of the compensation effect. The signal after compensation significantly improves the continuity of the current and eliminates the original current steps and waveform distortion. The harmonic content of the superimposed signal is measured to observe the suppression effect of the dead-time harmonics. In the traction system of a magnetic levitation train, dead-time compensation can significantly improve the torque ripple of the motor and improve ride comfort. By comparing and analyzing the total harmonic distortion, the improvement in signal quality before and after compensation is quantitatively evaluated. The compensated signal is re-evaluated for saturation state, and the improved evaluation criteria are used to distinguish between saturated and linear working states. When the harmonic distortion of the compensated signal is below the set threshold and the current amplitude is within the linear range, a linear state signal is generated. When the signal still has obvious nonlinear characteristics or the amplitude exceeds the saturation threshold, a saturated state signal is generated. After dead-time compensation, the electromagnet that was originally misjudged as a saturated state due to dead-time effects may be re-evaluated as a linear state, improving the accuracy of state detection.

[0037] The linear state signal is used to correct the saturation state signal to form a corrected current signal. The reference characteristics in the linear state signal, including the ideal current waveform, phase relationship and amplitude ratio, are extracted as the reference standard for correction. By comparing and analyzing the differences between the saturation state signal and the linear reference, the nonlinear effects caused by magnetic field saturation, such as waveform distortion, phase shift and amplitude compression, are identified. In practical applications, the magnetic suspension motor often encounters magnetic field saturation problems during startup, and at this time the control accuracy needs to be restored through correction. The correction adopts the idea of pre-distortion compensation, and by superimposing a reverse compensation component on the saturation state signal, the adverse effects of magnetic field saturation are offset. The generation of the compensation signal adopts a combination of lookup table method and real-time interpolation, and the correction coefficients under different saturation degrees are stored in advance, and the interpolation calculation is performed according to the actual saturation state during operation. The correction process considers the influence of the hysteresis loop, and adopts different correction strategies for magnetization and demagnetization processes. Under the condition of high-frequency varying load, the hysteresis effect will cause a phase difference between the current and the magnetic flux, which needs to be compensated by phase compensation to improve the control performance. The current signal after correction restores the nearly linear characteristics, and the waveform quality is significantly improved, and the harmonic distortion rate is reduced to an acceptable range. The corrected current signal retains the original control information while eliminating the nonlinear errors caused by magnetic field saturation.

[0038] The corrected current signal is distributed by the position deviation parameter to generate a distribution current vector. The position deviation parameter extracted from the radial displacement signal contains the offset amount and offset direction information of the rotor in each direction. According to the actual spatial position of the rotor, the contribution weight of each axis electromagnet to position correction is calculated. In the magnetic suspension bearing system, the rotor may deviate from the center position due to unbalanced force or external impact, and at this time multiple electromagnets need to work together to pull the rotor back to the balanced position. The weight distribution algorithm considers the spatial layout of the electromagnets, the direction of magnetic force and the geometric relationship of the position deviation. For radial deviation, it is mainly corrected by X-axis and Y-axis electromagnets, and the weight distribution is determined according to the deviation direction. The weight calculation formula is W_x=cos(θ), W_y=sin(θ), where θ is the deviation angle, and W_x and W_y are the weight coefficients of X-axis and Y-axis respectively. For example, when the rotor deviates to the northeast, both X-axis and Y-axis electromagnets need to participate in control, but their contribution proportions are dynamically adjusted according to the deviation angle. The Z-axis electromagnet mainly undertakes the control task of axial force, and the weight distribution is relatively independent. The corrected current signal is distributed according to the calculated weight coefficients to generate current instructions for each axis. Finally, a distribution current vector containing three-axis current instructions is generated, and each component of the vector corresponds to an accurate control current value.

[0039] The magnetic field saturation compensation control based on the distribution current vector is implemented to obtain the multi-degree-of-freedom control response value. The current instruction in the distribution current vector is input to the current loop controller of each axis to realize accurate current tracking through closed-loop control. The current loop controller adopts a PI regulator structure and has good steady-state accuracy and dynamic response characteristics. The magnetic field saturation compensation mechanism is integrated in the control process, and the control strategy is dynamically adjusted according to the real-time detection of the magnetic field state. For example, when the maglev train passes a curve, the outer electromagnet bears a larger lateral force and is prone to saturation. At this time, the system will automatically implement saturation compensation control. For the electromagnet close to saturation, current limiting and phase compensation measures are adopted to prevent control performance deterioration caused by deep saturation. The compensation control algorithm monitors the working state of each axis electromagnet, and when saturation trend is detected, the current instruction of the axis is automatically reduced, and the control components of other axes are increased to maintain the overall control effect. By measuring the output of each axis electromagnet in real time, the synthesis control force on the rotor is calculated. In the magnetic suspension flywheel energy storage system, the centrifugal force and gyroscopic force on the rotor in the high-speed rotating state will affect the suspension control, and the control parameters need to be optimized through force synthesis analysis. The size and direction of the control force reflect the control ability and response characteristics of the magnetic suspension system. Based on the analysis results of the control force, the multi-degree-of-freedom control response value is calculated, including key indicators such as radial stiffness, radial damping, axial stiffness, and angular stability.

[0040] In step S120, the multi-degree-of-freedom control response value is used to construct a decoupling control sequence, the stability of the decoupling control sequence is detected to obtain different axis data, the coupling coefficient sample set is extracted through inter-axis coupling analysis based on the different axis data, and the control parameter coordination matrix is constructed based on the coupling coefficient sample set.

[0041] Specifically, a decoupling control sequence is constructed using multi-degree-of-freedom control response values. According to the characteristic differences of the control response values of each axis, independent control channels are designed so that the control signals of each axis do not interfere with each other. For example, in a magnetic levitation precision machining spindle, the X-axis and Y-axis are responsible for radial position control, the Z-axis is responsible for axial position control, and the tilt axis is responsible for angle attitude control, and each axis needs to work independently to avoid mutual influence. The coupled multi-axis control problem is converted into multiple single-axis control problems through a decoupling transformation matrix, and the elements of the transformation matrix are determined according to the numerical relationship of the control response values. When constructing the decoupling control sequence, the control instructions of each axis are arranged in time sequence to ensure the coordination and timing of the control actions. In the suspension control of a magnetic levitation train, the control sequence of each electromagnet needs to be accurately synchronized to avoid instability caused by timing disorder. The decoupling control sequence contains the control amplitude, phase, and frequency information of each axis at each time, forming a complete set of multi-axis control instructions. The control sequence is sorted according to the execution priority, and the control instructions of the key axis have a higher execution priority. The decoupled control sequence eliminates the direct coupling relationship between the axes, and each axis can independently respond to its own control instructions, improving the control accuracy and response speed of the system.

[0042] The decoupling control sequence is subjected to stability detection to obtain different axis data. The constructed decoupling control sequence is input into the axis controllers to monitor the actual response and stability performance of each axis. By real-time acquisition of position feedback, speed feedback, and control force feedback of each axis, complete axis response data is obtained. In a magnetic levitation flywheel energy storage system, the dynamic characteristics of each axis are very different under high-speed rotation of the rotor, and the stability boundaries and working ranges of each axis need to be detected respectively. The stability detection uses amplitude-frequency characteristic analysis method, inputs a specific frequency excitation signal to each axis, and measures the corresponding response amplitude and phase change. During the detection process, the complete frequency band from low frequency to high frequency is scanned to identify the resonance frequency and anti-resonance frequency points of each axis. For each frequency point, the gain and phase margin of the axis are recorded to judge the stability degree of the system at that frequency. For example, in the radial control of a magnetic bearing system, the X-axis may resonate at a certain frequency, while the Y-axis remains stable at the same frequency. The frequency response function of each axis is obtained through frequency domain analysis, which reflects the dynamic transfer characteristics from control input to position output. The detected different axis data is classified and stored, including X-axis data, Y-axis data, Z-axis data, and angle axis data, each type of data containing the amplitude-frequency characteristic, phase-frequency characteristic, and stability index of the axis.

[0043] In some embodiments, the extracting the coupling coefficient sample set by the inter-axis coupling analysis on the different axial data comprises: detecting and identifying a dominant control axis, an auxiliary control axis and a balance control axis for the different axial data; establishing a control signal transmission path according to a control coordination relationship of the dominant control axis and the auxiliary control axis; generating a buffer adjustment parameter by adjusting a control parameter with the balance control axis; and extracting the coupling coefficient sample set based on the buffer adjustment parameter and the control signal transmission path.

[0044] The control state detection is performed on the different axial data to identify the dominant control axis, the auxiliary control axis and the balance control axis. The roles and positions of the axes in system control are determined by analyzing the control response strength, adjustment range and influence degree of each axial data. The dominant control axis is the axis with the greatest impact on system performance and the strongest control ability, which usually undertakes the main position adjustment and force output task. For example, in the radial control of a magnetic suspension bearing, the axis with the largest carrying capacity usually becomes the dominant control axis, which is responsible for the main radial support force adjustment. The axis with the largest control weight is identified by comparing the control gain, bandwidth and output power of each axis. The auxiliary control axis is the axis that assists the dominant control axis to complete the control task, which has certain control ability but relatively small influence. In the lateral stability control of a magnetic levitation train, the dominant control axis is responsible for the main lateral force adjustment, and the auxiliary control axis provides supplementary stability force. The balance control axis mainly plays a system balancing and fine tuning role, and optimizes the overall performance of the system through fine adjustment. The control characteristics of each axis are distinguished by comparative analysis of response speed, adjustment accuracy and steady-state error. The dominant control axis usually has a large adjustment range but relatively low accuracy, the auxiliary control axis has moderate adjustment ability and accuracy, and the balance control axis has high accuracy but small adjustment range.

[0045] The control coordination relationship between the dominant control axis and the auxiliary control axis is established to establish a control signal transmission path. According to the identified dominant control axis and auxiliary control axis, a coordinated control strategy and signal transmission mechanism between the two is designed. The dominant control axis serves as the main source of control instructions, delivering control information and coordination instructions to the auxiliary control axis. For example, in the XY plane control of a magnetic levitation precision positioning table, when the X axis is the dominant control axis, it needs to deliver speed information and position instructions to the Y axis to ensure coordinated movement of the two axes. The control signal transmission path includes two forms: direct transmission path and indirect transmission path. The direct transmission path is that the dominant control axis directly sends control signals to the auxiliary control axis, with minimal transmission delay but relatively low flexibility. The indirect transmission path transmits signals through a central coordinator for signal transfer and processing, which can perform more complex coordination operations but has slightly larger transmission delay. In the attitude control of a magnetic levitation flywheel, the control signals of the radial dominant axis need to be transmitted to other auxiliary axes after coordinate transformation. The bandwidth and delay characteristics of the signal transmission path directly affect the effect of coordinated control, which needs to be optimized according to the real-time requirements of the system.

[0046] The control parameter adjustment buffer is generated by means of the balance control axis. The high-precision adjustment capability of the balance control axis is used to finely adjust and buffer the control parameters of the dominant control axis and the auxiliary control axis. The balance control axis acts as a fine tuner and buffer for the system, improving overall control performance through small-scale parameter adjustment. For example, in a magnetic levitation centrifugal separator, the dominant control axis is responsible for the main radial force control, the auxiliary control axis provides supplementary support, and the balance control axis eliminates residual vibration through small force adjustment. The principle of buffer adjustment is to alleviate the impact of parameter changes on the system through reverse adjustment of the balance control axis when the control parameters change suddenly. When the output of the dominant control axis changes significantly, the balance control axis simultaneously performs small-scale reverse adjustment to reduce the transient response of the system. Buffer adjustment parameters include buffer gain, buffer time constant, and buffer range, etc. Buffer gain determines the response strength of the balance control axis to the changes of the dominant control axis. If the gain is too large, it may cause system oscillation, and if the gain is too small, the buffer effect will not be obvious. In the machining process of the main shaft of a magnetic levitation machine tool, the change of cutting force will affect the output of the dominant control axis, and the balance control axis will adjust the buffer to maintain machining precision.

[0047] Based on the buffer adjustment parameters and the control signal transmission path, a coupling coefficient sample set is extracted. The buffer adjustment parameters are taken as the excitation signal for system identification, and the parameter changes are transmitted to each control axis through the control signal transmission path. The response of each axis to the buffer adjustment parameter changes is monitored, and the amplitude, phase, and time delay characteristics of the response are measured. For example, in a magnetic levitation suspension system, by changing the buffer parameters of the balance control axis, the response changes in other axes are observed to identify the coupling relationship between the axes. Parameter identification algorithms such as least squares method and recursive identification method are used to extract the coupling coefficients from the input and output data. During the identification process, the delay and attenuation characteristics of the signal transmission path are considered, and the measured data are compensated and corrected accordingly. By changing the amplitude, frequency, and waveform of the buffer adjustment parameters, the coupling coefficient values under different conditions are obtained. The coupling coefficients obtained from multiple identifications are statistically analyzed to calculate the mean, variance, and confidence interval. The identification process is repeated under different load conditions, different speeds, and different environmental temperatures of the magnetic suspension system to obtain the variation law of the coupling coefficients. The extracted coupling coefficient sample set contains the coupling characteristics of the system under various working conditions, and the data quantity is rich and representative.

[0048] Based on the coupling coefficient sample set, a control parameter coordination matrix is constructed. The purpose of the coordination matrix is to compensate for the mutual interference caused by the coupling between the axes while ensuring the independent control ability of each axis. The rows and columns of the matrix correspond to different control axes, and the matrix elements represent the strength of the coordination relationship between the axes. For example, in a six-degree-of-freedom magnetic levitation platform, the coordination matrix is a 6×6 square matrix, and the diagonal elements represent the self-control strength of each axis, and the non-diagonal elements represent the cross-control strength between the axes. According to the statistical characteristics of the coupling coefficient sample set, the values of the elements of the coordination matrix are calculated. The calculation of the matrix elements considers multiple factors such as coupling strength, control priority, and system stability. For strongly coupled axial combinations, the corresponding coordination coefficients are increased to strengthen the coordination control between the axes. For weakly coupled or unrelated axial combinations, the coordination coefficients are set to small values or zero. In the control of the bogie of a magnetic levitation train, the suspension control on the left and right sides needs to be strongly coordinated, while the coordination between the front and rear axes requires relatively low coordination. The finally constructed control parameter coordination matrix can achieve optimal coordination of multi-axis control, ensuring control accuracy and maintaining system stability.

[0049] In step S130, control efficiency characteristic analysis is performed on the control parameter coordination matrix to identify an efficient control area, load characteristics are identified based on multi-axis current feedback signals to determine a current load mode, a mode matching working point position is selected from the efficient control area according to the current load mode, and a coordination control node is determined by comparing the mode matching working point position with multi-degree-of-freedom control response values.

[0050] Specifically, the control parameter coordination matrix is analyzed to identify the high-efficiency control area. The coordination matrix reflects the interaction relationship between the axes in the multi-axis control system, and different matrix element combinations correspond to different control efficiencies. By calculating the eigenvalues and condition numbers of the matrix, the numerical stability and convergence characteristics of the coordination control are evaluated. For example, in a magnetic levitation precision positioning platform, the coordination control matrix of the XY axis has multiple working areas, some of which can achieve faster response speed and higher positioning accuracy. The efficiency characteristic analysis uses the energy function method to calculate the control energy consumption and response time under different matrix parameter combinations. By scanning the parameter space of the coordination matrix, the contour map of control efficiency is drawn, which intuitively shows the trend and rule of efficiency change. The high-efficiency control area corresponds to the area with dense efficiency contours and large numerical values. These areas have high control gain and low energy loss. In a magnetic levitation flywheel energy storage system, the optimal coordination matrix parameters change at different speeds, and the high-efficiency control area needs to be identified at each speed section. Through frequency domain analysis method, the transfer function amplitude of the coordination matrix at different frequencies is calculated to identify the parameter interval with good frequency response characteristics. The identified high-efficiency control areas are classified according to the efficiency level to form a multi-level efficiency area division.

[0051] In some embodiments, the load feature recognition based on the multi-axis current feedback signal determines the current load mode, including: establishing a load disturbance detection using the multi-axis current feedback signal to generate disturbance feature data; evaluating the disturbance amplitude in the disturbance feature data to find the load mutation degree; performing hierarchical control processing on the load mutation degree to form a buffer control sequence; and matching and identifying the current load mode according to the buffer control sequence.

[0052] For example, the establishment of load disturbance detection using the multi-axis current feedback signal to generate disturbance feature data includes: tracking the cross-cycle changes of the multi-axis current feedback signal to obtain change evolution data; analyzing the disturbance response amount of each axis load based on the change evolution data to establish disturbance quantization indicators, including radial load, axial load and torque load; and mapping the disturbance quantization indicators to generate disturbance feature data.

[0053] The multi-axis current feedback signals are tracked for cross-cycle variation to obtain variation evolution data. By continuously monitoring the variation of multi-axis current feedback signals in multiple working cycles, the development and evolution process of load disturbance and long-term change trend are tracked. Cross-cycle variation tracking can identify slow developing load changes and periodic load patterns that may not be obvious in a single cycle but will be revealed under the cumulative effect of multiple cycles. For example, in the long-term operation of magnetic suspension bearings, the slow change of bearing gap will cause the drift of current reference value, which needs to be identified through cross-cycle tracking. The variation tracking adopts a moving reference method, using the average value of multiple historical cycles as the reference to calculate the deviation of the current cycle relative to the reference. By establishing a time series model of current variation, different components such as trend items, seasonal items and random items in the signal are identified. In the daily operation of magnetic levitation trains, the periodic variation of load caused by passenger flow changes in different time periods needs to be identified through cross-cycle analysis. The tracking process records the key feature points of the current signal in each cycle, including peak value, mean value, variance and spectral barycenter, etc. By comparing the changes of adjacent cycle feature points, the change rate and change direction are calculated. The variation evolution data contains the complete historical information of load variation, reflecting the dynamic evolution process of system load state.

[0054] Based on the variation evolution data analysis of the disturbance response of each axial load, a disturbance quantification index is established. The current amplitude variation ΔI_cycle and the change rate dI / dt of each working cycle are extracted from the variation evolution data as the basic data for analysis of the disturbance of each axial load. The radial load disturbance response is calculated by analyzing the current fluctuations of X and Y axes in the variation evolution data, and the radial disturbance intensity R_radial=√((ΔI_X)²+(ΔI_Y)²) is calculated using the current variation amplitude in the evolution data, where ΔI_X is the X-axis current variation amplitude and ΔI_Y is the Y-axis current variation amplitude. The axial load disturbance response is determined by analyzing the current variation characteristics of the Z-axis electromagnet in the variation evolution data, and the axial disturbance intensity A_axial=|ΔI_Z_max-ΔI_Z_min| is calculated based on the periodic variation of Z-axis current in the evolution data, where ΔI_Z_max is the maximum value of Z-axis current variation and ΔI_Z_min is the minimum value of Z-axis current variation. The torque load disturbance response reflects the change of the phase relationship of each axis electromagnet current, and the torque disturbance intensity T_torque=Σ|Δφ_i| is calculated by analyzing the phase shift angle φ of each axis current in the variation evolution data, where Δφ_i is the phase shift angle of the i-th axis and Σ represents the summation of all axes. The radial, axial and torque disturbance response quantities calculated based on the variation evolution data are comprehensively processed to establish a disturbance quantification index system.

[0055] The disturbance quantization indicators are mapped to feature distribution to generate disturbance feature data. By projecting the established disturbance quantization indicators into a multi-dimensional feature space, a distribution map and data set of disturbance features are formed. Dimension reduction techniques such as principal component analysis are used to map the high-dimensional quantization indicator space to a two-dimensional or three-dimensional visualization space. For example, in the application of a magnetic levitation magnetic force pump, the disturbance features under different fluid working conditions form different clustering regions in the mapping space, facilitating working condition recognition and control strategy selection. The feature distribution mapping also uses kernel density estimation methods to calculate the probability density distribution of the disturbance features in the space, identifying high-density regions and boundary regions. The data points in the mapping space are divided into several disturbance categories by clustering algorithms, and each category corresponds to a specific disturbance mode. In the operation of a magnetic levitation high-speed motor, the disturbance features in different speed ranges exhibit different distribution patterns in the mapping space. The mapping results also contain weight information of each feature dimension, reflecting the contribution of different quantization indicators to disturbance classification. The generated disturbance feature data has a structured format, containing feature vectors, category labels, and confidence information.

[0056] The disturbance amplitude evaluation in the disturbance feature data finds the degree of load mutation. By analyzing the amplitude information in the disturbance feature data, the mutation degree and influence range of the load change are evaluated. The mutation degree of the load reflects the severity of the system disturbance and is an important basis for selecting control strategies. The disturbance amplitude evaluation uses a multi-level evaluation system to divide the mutation degree into five levels: slight, slight, moderate, severe, and extremely severe. For example, in a magnetic levitation precision machining device, the mutation of cutting force will affect the machining precision, and the corresponding compensation control strategy needs to be selected according to the mutation degree. The evaluation process first calculates the percentage of the disturbance amplitude relative to the normal working current as the basic indicator of mutation intensity. By statistical analysis of historical disturbance data, the classification threshold and evaluation standard of mutation degree are established. For multi-axis systems, the disturbance amplitudes of each axis need to be considered, and the overall mutation degree is determined by weighted summation or maximum value. In a magnetic levitation flywheel system, the radial disturbance and axial disturbance have different effects on system stability, and different weight coefficients need to be used. The disturbance amplitude evaluation also considers the change rate of the disturbance, and the fast-changing disturbance is more likely to cause system instability than the slow-changing disturbance.

[0057] The load mutation degree is classified and controlled to form a buffer control sequence. According to the evaluated load mutation degree, a corresponding level of control response strategy is designed, and a systematic buffer control mechanism is formed through hierarchical processing. Different levels of load mutation require different intensity and speed of control response to ensure the stability and control accuracy of the system. For example, in a magnetic levitation elevator system, the load change caused by passengers getting on and off the elevator belongs to moderate mutation, and gradual control adjustment is needed to ensure ride comfort. The hierarchical control processing first selects the control gain and response speed according to the mutation degree, and small gain and slow response for slight mutation, and large gain and fast response for severe mutation. The core idea of buffer control is to gradually adjust the control parameters to alleviate the impact of mutation on the system when the load mutates. For minor mutations, the system maintains the current control parameters unchanged, relying on the system's own robustness to adapt to changes. For slight and moderate mutations, the buffer control mechanism is started, and the control parameters are gradually adjusted according to the preset time sequence. In a magnetic levitation magnetic stirrer, changes in liquid viscosity will cause load mutation, which requires buffer control to maintain the stability of the stirring speed. The buffer control sequence contains multiple control stages, each stage corresponding to a specific control parameter combination and duration.

[0058] The current load mode is identified according to the buffer control sequence matching. Different load modes correspond to specific buffer control sequence modes, and the current load state can be accurately identified through sequence matching. The load mode library contains standard buffer control sequences of the system under various typical working conditions, and each mode has a clear feature description and application range. For example, in a magnetic levitation wind turbine, the load characteristics under different wind speed and direction conditions differ greatly, corresponding to different load modes and buffer control sequences. The matching identification uses a pattern recognition algorithm to calculate the similarity of the current buffer control sequence with each standard mode. The similarity calculation considers multiple dimensions such as amplitude characteristics, time sequence characteristics, and frequency characteristics. By setting a matching threshold, when the similarity exceeds the threshold, it is confirmed that the matching is successful, and the corresponding load mode is identified. In complex working conditions where multiple modes coexist, multiple load modes may be activated simultaneously, and the dominant mode needs to be determined through weighted fusion. The matching process also considers the time correlation of the sequence, and the sequence trend in consecutive time periods is also an important feature of pattern recognition.

[0059] According to the current load mode, a mode matching working point position is selected from the high-efficiency control area. Different load modes require corresponding control strategies, and there are optimal working points in the high-efficiency control area that match various load modes. For example, when the magnetic suspension bearing bears impact load, a working point with fast response characteristics needs to be selected, and when in steady state operation, a working point with lower energy consumption is preferred. The working point selection algorithm searches for the parameter combination with the highest matching degree in the high-efficiency control area according to the characteristic parameters of the load mode. The matching degree calculation considers multiple factors such as load intensity, frequency characteristics and duration, and quantifies the matching degree through weighted scoring. In the application of the magnetic suspension magnetic force pump, the hydraulic load characteristics under different flow conditions differ greatly, and the corresponding optimal working point needs to be selected to ensure system efficiency. The selection process adopts a multi-objective optimization strategy to find a balance point between control accuracy, response speed and energy efficiency. Through intelligent optimization methods such as genetic algorithm or particle swarm algorithm, the optimal working point position is quickly converged. The selected working point position contains complete information such as the specific parameter value of the coordination matrix, the control gain setting and the frequency response characteristics.

[0060] In some embodiments, the determining a coordinated control node by comparing the mode matching working point position with the multi-degree-of-freedom control response value comprises: obtaining a magnetic field loss characteristic of each frequency component in the mode matching working point position; matching and comparing the multi-degree-of-freedom control response value with the magnetic field loss characteristic to generate a comparison result; performing PWM control frequency optimization based on the comparison result to generate optimal control parameters; and constructing a coordinated control strategy according to the optimal control parameters to determine a coordinated control node.

[0061] The magnetic field loss characteristics of each frequency component in the mode matching operating point position are obtained. By analyzing the control parameter combination corresponding to the mode matching operating point position, the magnetic field loss distribution and characteristics in the electromagnet core and winding under different frequency components are calculated. The magnetic field loss includes two main parts: core loss and copper loss. The loss characteristics directly affect the efficiency and temperature rise of the system. For example, in the application of magnetic levitation high-speed motors, different PWM frequencies will result in different magnetic field loss modes, and the loss contribution of each frequency component needs to be analyzed. The core loss is mainly composed of hysteresis loss and eddy current loss, where the hysteresis loss is proportional to the magnetization frequency, and the eddy current loss is proportional to the square of the frequency. Through finite element analysis method, the magnetic field distribution and magnetic flux density variation corresponding to the operating point position are calculated. At different frequencies, the penetration depth and distribution mode of the magnetic field will change, affecting the spatial distribution of the loss. The copper loss is mainly composed of winding resistance loss and alternating current resistance loss, and high-frequency components will increase the alternating current resistance of the conductor. By spectral analysis, the current spectral components corresponding to the operating point position are extracted, and the effective value and phase relationship of each frequency component are calculated. The magnetic field loss characteristics include total loss, loss density distribution, frequency response, and temperature coefficient, etc. In the magnetic levitation bearing system, the loss characteristics of radial electromagnets and axial electromagnets are quite different, and need to be analyzed separately.

[0062] The multi-degree-of-freedom control response value is matched with the magnetic field loss characteristics to generate a comparison result. By comparing and analyzing the relationship between the multi-degree-of-freedom control response value and the magnetic field loss characteristics, the matching degree of control performance and energy efficiency is evaluated. The comparison process needs to find the best balance point between control accuracy, response speed and energy efficiency. For example, in the magnetic levitation precision positioning system, high-precision control usually requires high control bandwidth and PWM frequency, but this will increase the magnetic field loss. The comparison analysis uses a multi-objective evaluation function to consider the weighted combination of control performance indicators and loss indicators. Through normalization processing, different dimension indicators are converted into a unified evaluation standard. The radial stiffness, axial stiffness and angle stability in the control response value are compared one by one with the corresponding magnetic field loss. In the magnetic levitation flywheel energy storage system, high radial stiffness is needed to suppress vibration at high speed, but too high stiffness will increase the loss of the control system. The comparison result is visualized in the form of radar chart or scatter plot, directly reflecting the matching of each indicator. Through cluster analysis, the advantage area and disadvantage area in the comparison result are identified, and the advantage area corresponds to the operating point with good control performance and low loss.

[0063] Based on the comparison result, the PWM control frequency optimization is generated to generate the optimal control parameters. The matching degree score and loss distribution information in the comparison result are extracted, and the PWM frequency is adjusted for the working point with a matching degree lower than 0.7. The comparison result shows that the current PWM frequency has a high magnetic field loss in the high frequency band, and the switching frequency needs to be reduced to improve the efficiency characteristics. The loss-frequency relationship curve in the comparison result is used to determine that the optimal frequency range is between 8-12 kHz. For example, in the magnetic suspension bearing application, the comparison result shows that the PWM frequency of 15 kHz leads to a high core loss, and the total loss can be reduced by 30% by adjusting the frequency to 10 kHz. According to the matching of each axis in the comparison result, different optimal frequencies are selected for the X-axis, Y-axis, and Z-axis. The X-axis has a good comparison result, and the frequency is maintained at 12 kHz; the Y-axis has a high loss in the comparison result, and the frequency is adjusted to 9 kHz; the Z-axis has an average comparison result, and the frequency is set to 10.5 kHz. The control accuracy requirement in the comparison result is weighed with the frequency optimization to ensure that the frequency reduction does not affect the control bandwidth. Through the radar chart analysis of the comparison result, the best balance point of control performance and loss is identified. The generated optimal control parameters include the PWM frequency of each axis, the corresponding control gain compensation coefficient, and the dead time adjustment value.

[0064] According to the optimal control parameters, a coordinated control strategy is constructed to determine the coordinated control node. The PWM frequency in the generated optimal control parameters is combined as the core configuration of the coordinated strategy. Based on the different PWM frequency settings of each axis, a corresponding synchronous coordination mechanism is designed to ensure that the control instructions under different frequencies can be executed in coordination. For example, in the magnetic suspension precision positioning table, the 4:3 frequency ratio relationship between the 12 kHz frequency of the X-axis and the 9 kHz frequency of the Y-axis needs to be maintained stable in the coordination strategy. The gain compensation coefficient in the optimal control parameters is used to adjust the signal amplification multiple of each axis control node, with the X-axis gain coefficient being 1.0, the Y-axis gain coefficient being 1.15, and the Z-axis gain coefficient being 1.08. According to the dead time setting in the optimal control parameters, the timing control logic of each coordinated control node is configured. The X-axis with the highest frequency in the optimal control parameters is set as the master control node, which is responsible for generating a synchronous reference signal; the Y-axis and the Z-axis with lower frequencies are set as slave control nodes, which are synchronized according to the reference signal of the master node. The communication protocol of the coordinated control node is configured according to the timing requirements of the optimal control parameters, and the master node sends a synchronization pulse and control instructions to the slave node every control cycle. Based on the frequency allocation of the optimal control parameters, the data transmission timing between nodes is determined to avoid interference between nodes working at different frequencies.

[0065] In step S140, the control weight distribution correction is performed according to the coordinated control node to determine the conversion parameter set. The multi-axis coordinated sequence is composed of the power switch controller through the conversion parameters, and the multi-axis coordinated control strategy is generated by synchronously generating the multi-axis PWM signals of the multi-axis coordinated sequence.

[0066] In some embodiments, the control weight distribution correction determination of the transition parameter set according to the coordination control node comprises: generating a control demand distribution for each axis control demand level evaluation of the coordination control node; dividing the controller into a high response control area and an energy saving control area based on the control demand distribution; switching the controller of the energy saving control area to a light load working mode to obtain an energy saving control signal; and determining the transition parameter set in coordination with the high response control area according to the energy saving control signal.

[0067] The control demand distribution is generated for each axis control demand level evaluation of the coordination control node. The control ability and actual load demand of the master node and the slave node in the coordination control node are analyzed, and the control demand level of each axis is quantitatively evaluated. The PWM frequency of the X-axis of the master coordination control node is the highest (12 kHz) and the gain coefficient is the reference value (1.0), indicating that the control task of this axis is the heaviest, and the control demand level is rated as high, requiring fast response and high precision control ability. The gain coefficient of the Y-axis slave coordination control node is the highest (1.15), indicating that a larger control compensation is needed, but the PWM frequency is relatively low (9 kHz), and the control demand level is evaluated as medium, requiring medium response speed and control accuracy. The PWM frequency (10.5 kHz) and gain coefficient (1.08) of the Z-axis slave coordination control node are both at a medium level, and the control demand level is rated as medium-low, requiring relatively low response speed and control accuracy. Through the frequency-gain combination characteristics of the coordination control node, an evaluation model of the control demand level is established. High frequency and high gain correspond to high control demand, low frequency and low gain correspond to low control demand, and medium parameters correspond to medium control demand. For example, in the magnetic levitation train suspension system, the master bearing coordination control node needs high-level control to quickly respond to track excitation and load changes, and the auxiliary stability coordination control node only needs medium-level control to maintain system balance. According to the control demand level of each coordination control node, a control demand distribution map is generated.

[0068] The controller is divided into high-response control area and energy-saving control area based on control demand distribution. The generated control demand distribution information is used to distribute the power controller of each axis to different working areas according to the demand intensity and importance. The X-axis in the control demand distribution belongs to the high demand level, and the corresponding power controller is divided into the high-response control area. The controller in this area needs to maintain the best dynamic response performance and control accuracy. The Y-axis and Z-axis in the control demand distribution belong to the medium and low levels, and the corresponding power controller is divided into the energy-saving control area. The controller in this area can optimize the energy efficiency performance on the premise of meeting the basic control requirements. The division of the high-response control area and the energy-saving control area reflects the rational allocation of system resources. The key control task is given priority protection, and the secondary control task focuses on energy efficiency. For example, in the magnetic levitation flywheel energy storage system, the power controller corresponding to the radial main coordination control node needs to bear the main radial load, and is divided into the high-response control area to ensure the rapid response ability. The controller corresponding to the auxiliary coordination control node can work in the energy-saving control area to reduce the total power consumption of the system. The power controller in the high-response control area maintains the high-frequency PWM working mode and the maximum control gain to ensure sufficient control ability under various working conditions.

[0069] The controller in the energy-saving control area is switched to the light-load working mode to obtain the energy-saving control signal. The Y-axis power controller and the Z-axis power controller divided into the energy-saving control area are converted to the working mode. The Y-axis power controller is switched from the standard working mode to the light-load working mode. The PWM switching frequency is appropriately reduced from 9 kHz to reduce the switching loss. The power output level is adjusted according to the light-load demand. The control gain coefficient is adjusted from 1.15 to a value suitable for the light-load working condition. The Z-axis power controller also implements light-load working mode switching. The working frequency is reduced from 10.5 kHz. The power output and gain parameters are optimized and adjusted according to the energy-saving requirements. The design principle of the light-load working mode is to maximize the energy-saving effect on the premise of ensuring the basic control function. The controller can still maintain the necessary position adjustment and force output ability, but the response speed and power consumption are significantly reduced. The Y-axis power controller in the light-load mode outputs the energy-saving control signal, which contains the degraded power command, the adjusted switching frequency information, and the synchronization information required for coordination with other controllers. The Z-axis power controller in the light-load mode also outputs the corresponding energy-saving control signal, which reflects the working state and control ability under the light-load working condition. For example, in the magnetic levitation spindle system of the machine tool, the power controller of the non-key axis electromagnet can work in the light-load mode to effectively reduce the total power consumption of the system on the premise of ensuring the machining accuracy. Through the implementation of the light-load working mode, the energy-saving control area realizes the optimization balance of power consumption and control performance, and provides technical support for the efficient operation of the system.

[0070] The conversion parameter set is generated according to the energy-saving control signal and the high-response control area. The energy-saving control signal output by the energy-saving control area contains the working parameter information of the Y-axis and Z-axis power controllers in the light load mode, and is coordinated with the X-axis power controller of the high-response control area. The energy-saving control signal shows that the output capacity of the Y-axis and Z-axis power controllers has decreased, and the control tasks and power sharing ratio among the three axes need to be redistributed. According to the power degradation information in the energy-saving control signal, the power redistribution parameter Px_new=Px_base+(Py_loss+Pz_loss)×kx is calculated, where Px_new is the redistributed power of the X-axis, Px_base is the basic power of the X-axis, Py_loss and Pz_loss are the power loss of the Y-axis and Z-axis, and kx is the power distribution coefficient of the X-axis. The frequency adjustment information in the energy-saving control signal generates the timing conversion parameter to ensure that the controllers of each axis remain in a synchronous and coordinated relationship after the frequency is reduced. Based on the gain adjustment in the energy-saving control signal, the weight correction parameter Wc_i=Gx_ref / Gi_eco is calculated, where Wc_i is the weight correction coefficient of the i-axis, Gx_ref is the reference gain of the X-axis, and Gi_eco is the gain of the i-axis in the energy-saving mode. The power redistribution parameter, timing conversion parameter and weight correction parameter are arranged and combined in the axial order to generate the conversion parameter set: the X-axis conversion parameter contains the power increment and the reference timing, the Y-axis conversion parameter contains the power degradation and the timing delay, and the Z-axis conversion parameter contains the power adjustment and the corresponding delay.

[0071] The multi-axis coordination sequence is composed of power switch controllers using the conversion parameter set. The conversion parameter set is input into the corresponding power switch controller for configuration. The X-axis power switch controller receives the power increment information in the conversion parameter, sets the dominant working state according to the reference timing, and undertakes additional control tasks to compensate for the capacity decrease of other axes. The Y-axis power switch controller adjusts the output power level according to the power degradation amount in the conversion parameter, and adjusts the start time of the switch action according to the timing delay parameter. The Z-axis power switch controller sets the output level according to the power adjustment amount in the conversion parameter, and performs timing coordination according to the corresponding delay parameter. For example, in the application of magnetic suspension centrifuge, the three-axis power switch controllers are started in turn according to the configuration requirements of the conversion parameter, and the X-axis controller is started first to undertake the main control task, and the Y-axis and Z-axis controllers are started in turn according to their respective delay parameters. The on-off sequence and power distribution relationship of each power switch controller are accurately controlled through the timing configuration information in the conversion parameter. The three power switch controllers configured are combined into a multi-axis coordination sequence according to the power distribution and timing relationship determined by the conversion parameter. The coordination sequence is executed in order according to the configuration logic of the conversion parameter, and the accurate coordination of the three-axis power output is realized.

[0072] The multi-axis PWM signal synchronization generation multi-axis coordination sequence is a multi-axis coordination control strategy. The PWM signals output by the three power switch controllers in the multi-axis coordination sequence are uniformly synchronized and coordinated. The X-axis controller outputs a 12 kHz PWM signal, the Y-axis controller outputs a 9 kHz PWM signal, and the Z-axis controller outputs a 10.5 kHz PWM signal in the multi-axis coordination sequence. The three different frequency signals need to be coordinated through a synchronization mechanism. The timing relationship of the multi-axis coordination sequence is used to establish a unified common clock reference, and each axis PWM signal generates a synchronous pulse according to the corresponding frequency division relationship. Through frequency relationship calculation, the least common multiple of the three frequencies is determined as the main clock frequency, and the X-axis, Y-axis and Z-axis generate their own PWM signals according to the corresponding frequency division ratio. The power distribution relationship in the multi-axis coordination sequence directly affects the duty cycle setting of the PWM signal of each axis. According to the weight correction coefficient in the conversion parameter, the duty cycle of the X-axis PWM signal remains standard, the duty cycle of the Y-axis PWM signal is modulated according to the 0.87 coefficient, and the duty cycle of the Z-axis PWM signal is modulated according to the 0.93 coefficient. For example, in a magnetic suspension precision machining spindle, the synchronization accuracy of the three-axis PWM signal directly affects the rotation accuracy of the spindle and the machining surface quality, and the multi-axis coordination sequence ensures the accurate synchronization of the control signals of each axis. The three-axis PWM signals after synchronization are combined to form a complete multi-axis coordination control strategy. The coordination control strategy contains complete control parameter information such as PWM frequency setting, phase relationship, duty cycle distribution and synchronization timing of each axis.

[0073] In step S150, the dynamic characteristic data is obtained by analyzing the response speed of the multi-axis current feedback signal, and the control response period is determined based on the dynamic characteristic data and the multi-axis coordination control strategy. The multi-axis synchronization conversion operation is performed to generate the coordination control instruction according to the control response period.

[0074] In some embodiments, the dynamic characteristic data is obtained by analyzing the response speed of the multi-axis current feedback signal, including: detecting the current response time based on the multi-axis current feedback signal to obtain a response time component; analyzing the dynamic response characteristic of the response time component to establish a dynamic response parameter; generating a dynamic compensation control signal by dynamic control compensation processing according to the dynamic response parameter; and determining the dynamic characteristic data through the dynamic compensation control signal.

[0075] The current response time detection is based on multi-axis current feedback signals to obtain response time components. The current waveform data of X-axis, Y-axis and Z-axis in the multi-axis current feedback signals are received, and the response time characteristics of each axis current signal are separated and detected. The current response time detection adopts a pulse excitation method, and a standard pulse voltage signal is input to each axis to monitor the time course of the current rising from zero to the target value. The delay time td of the current response is measured, i.e. the time interval from the start of the input pulse to the start of the current rising, and the delay time reflects the pure lag characteristics of the system. The linear segment time tl of the current rising is detected, i.e. the time length of the current maintaining an approximate constant slope in the main rising stage, and the linear segment time reflects the main time constant of the system. The transition time tt of the current response is recorded, i.e. the time from the end of the linear rising to the steady-state value, and the transition time reflects the high-order dynamic characteristics of the system. For example, in the magnetic levitation train suspension system, the differences in the current response time of each suspension electromagnet will affect the coordinated allocation of the suspension force, and it is necessary to accurately detect the differences in the time components of each axis. Through the synchronous sampling of the multi-axis current feedback signals, the relative relationship and synchronization characteristics of the response time of each axis are detected. The fast component and the slow component in the current response of each axis are separated, the fast component corresponds to the inductance response of the circuit, and the slow component corresponds to the magnetization response of the magnetic circuit. The delay time, the linear segment time and the transition time detected are combined to form a complete set of response time components of each axis.

[0076] The dynamic response parameters are established by dynamic response characteristic analysis of the response time components. Through the combination relationship of the response time components, the dynamic response types of each axis control system are identified, including different types such as first-order system, second-order system and high-order system. For the response characteristics with short delay time and long transition time, it is determined as a first-order system type, and the corresponding dynamic response parameter is time constant τ = tl + tt. For the response characteristics with obvious overshoot and oscillation, it is determined as a second-order system type, and the dynamic response parameters such as damping ratio ζ and natural frequency ωn are calculated. Through the frequency domain transformation of the response time components, the frequency response characteristics and bandwidth parameters of each axis system are analyzed. The mapping relationship between the dynamic response parameters and the response time components is established to form a parameterized dynamic model description. For example, in the magnetic levitation machine tool spindle system, the dynamic response characteristics of the radial bearing and the axial bearing are quite different, and the corresponding dynamic response parameters need to be established respectively. The dynamic response parameters of each axis are compared and analyzed to identify the influence degree of the parameter differences on the multi-axis coordinated control.

[0077] The dynamic compensation control signal is generated according to the dynamic response parameter. Based on the established dynamic response parameters, including time constant, damping ratio, natural frequency and other key parameters, the corresponding dynamic control compensation strategy is designed. The purpose of dynamic control compensation is to improve the dynamic response performance of the system through prediction and advance adjustment, and to reduce the response delay and overshoot phenomenon. For the shaft with large time constant, the feedforward compensation method is adopted, and the control command is issued in advance according to the time constant τ, and the compensation time is 0.5τ. For the shaft with small damping ratio which is easy to oscillate, the damping compensation method is adopted, and the oscillation is suppressed by increasing the damping of the control loop. The dynamic control compensation process is realized by using digital filter, and the parameters of the filter are adjusted adaptively according to the dynamic response parameters. A lead correction link is designed, and the transfer function is Gc(s)=(1+aTd×s) / (1+Td×s), where Td is the lead time constant, and a is the lead proportional coefficient, which is determined according to the dynamic response parameters of each shaft. For example, in the magnetic suspension magnetic stirrer, the dynamic characteristics of each support point of the stirring shaft are different, and different compensation strategies are required to realize the coordinated control of each support point through dynamic compensation. The compensation process also considers the dynamic coupling effect between the shafts, and reduces the interference between the shafts through cross compensation. The generated dynamic compensation control signal contains the control command information of each shaft after compensation optimization.

[0078] The dynamic characteristic data is determined by the dynamic compensation control signal. The time domain response characteristics of each shaft are analyzed from the dynamic compensation control signal, and the dynamic time parameters are determined by measuring the rising process and stable process of the compensation signal. The rise time tr=(t90%-t10%) / 0.8 and the regulation time ts=t_settle of each shaft are calculated, where t90%, t10% are the times when the compensation signal reaches 90% and 10% amplitude, and t_settle is the signal settling time. The rise time reflects the fast response ability of the system after compensation, and the regulation time reflects the complete time characteristic of the system to reach steady state. Through frequency domain analysis of the dynamic compensation control signal, the frequency response characteristics and system bandwidth information of each shaft are extracted, and the frequency domain characteristics and time domain parameters are verified with each other to ensure the accuracy of the dynamic characteristic data. The response consistency of the dynamic compensation control signal under different working conditions is analyzed, and the stability and repeatability of the compensation effect are evaluated. The response speed and stability of the dynamic compensation control signal of each shaft are compared and evaluated, and the difference degree and mismatching quantitative index of the dynamic characteristics between the shafts are identified. The rise time and regulation time obtained from the dynamic compensation control signal analysis are arranged in the order of shaft, and a structured dynamic characteristic data matrix is formed, and the matrix elements directly correspond to the specific time characteristic values of each shaft.

[0079] The control response period is determined based on the dynamic characteristic data and the multi-axis coordination control strategy. The PWM frequency setting in the multi-axis coordination control strategy is matched and analyzed with the rise time tr and the regulation time ts in the dynamic characteristic data to ensure the coordination of the control period and the system dynamic characteristics. According to the rise time of each axis in the dynamic characteristic data, the lower limit value of the control response period is determined, and the control period must be greater than 2 times the maximum rise time to ensure that the control command has enough time to complete the response process. Based on the regulation time of each axis in the dynamic characteristic data, the upper limit value of the control response period is set, and the control period cannot exceed 1.5 times the minimum regulation time to ensure the real-time performance and response speed of the control. By comprehensively considering the time constraint and the frequency constraint, the control response period Tc is calculated as max(2×tr_max, PWM_period×N), where tr_max is the maximum rise time of each axis, PWM_period is the PWM period, and N is an integer multiple coefficient. The selection of the control response period needs to balance the control accuracy and the system stability, and a too short period may cause the control command to be updated too frequently and cause oscillation, and a too long period may reduce the dynamic response ability of the system. The determined control response period is matched and verified with the timing requirements of the multi-axis coordination control strategy to check the compatibility of the period setting and the synchronization relationship of the PWM signals of each axis, and to ensure the timing coordination and synchronization accuracy of the multi-axis coordination control.

[0080] The multi-axis synchronous conversion operation is performed according to the control response period to generate the coordinated control command. The synchronous conversion operation needs to complete the update, transmission and execution of the control command of each axis within the control response period to ensure the coordination and synchronization of multi-axis control. At the beginning of each control response period, the current state information of each axis is read simultaneously, including the current feedback value, the position feedback value and the control error, etc. According to the requirements of the multi-axis coordination control strategy, the target control amount of each axis in the next control period is calculated. The target control amount of each axis is converted into specific PWM control command through the synchronous conversion algorithm, and the coupling relationship and coordination requirements between axes are considered in the conversion process. For example, in the magnetic suspension flywheel energy storage system, the control commands of the radial axes need to maintain phase coordination to avoid generating unbalanced torque to affect the stable suspension of the rotor. The multi-axis synchronous conversion operation adopts a master-slave synchronization mode, the master axis sends a synchronization signal at the beginning of the control response period, and the slave axis receives the synchronization signal and executes the control command conversion at the same time. The execution time of the conversion operation is controlled to be completed within the first 1 / 3 of the control response period, and the remaining time is used for stable execution of the control command. The generated coordinated control command contains complete control information such as the PWM duty ratio, the switching timing, the power distribution and the protection parameters of each axis. The coordinated control command is sent to the power controller of each axis through a high-speed communication interface to realize accurate coordination control of multi-axis.

[0081] In order to implement the multi-axis coordinated control method of the magnetic suspension motor corresponding to the above-mentioned method embodiment, the corresponding functions and technical effects are realized. Figure 2 The structural block diagram of the multi-axis coordinated control system 200 of the magnetic suspension motor provided by the embodiment of the application is shown. For the convenience of description, only the parts related to the embodiment are shown. The multi-axis coordinated control system 200 of the magnetic suspension motor provided by the embodiment of the application comprises:

[0082] The signal acquisition module 201 is configured to acquire the multi-axis current feedback signal and the radial displacement signal of the magnetic suspension motor, perform multi-channel PWM modulation on the multi-axis current feedback signal to obtain a PWM modulated current signal, extract a position deviation parameter from the radial displacement signal, and couple and associate the PWM modulated current signal with the position deviation parameter to generate a multi-degree-of-freedom control response value.

[0083] The coupling analysis module 202 is configured to construct a decoupling control sequence by using the multi-degree-of-freedom control response value, perform stability detection on the decoupling control sequence to obtain different axial data, perform inter-axis coupling analysis by using the different axial data to extract a coupling coefficient sample set, and construct a control parameter coordination matrix based on the coupling coefficient sample set.

[0084] The load identification module 203 is configured to perform control efficiency characteristic analysis on the control parameter coordination matrix to identify an efficient control area, perform load characteristic identification based on the multi-axis current feedback signal to determine a current load mode, select a mode matching working point position from the efficient control area according to the current load mode, and determine a coordinated control node by comparing the mode matching working point position with the multi-degree-of-freedom control response value.

[0085] The control coordination module 204 is configured to perform control weight distribution correction to determine a conversion parameter set according to the coordinated control node, use the conversion parameter to form a multi-axis coordination sequence through a power switch controller, perform multi-axis PWM signal synchronization on the multi-axis coordination sequence to generate a multi-axis coordinated control strategy.

[0086] The dynamic adjustment module 205 is configured to perform response speed analysis on the multi-axis current feedback signal to obtain dynamic characteristic data, determine a control response period based on the dynamic characteristic data and the multi-axis coordinated control strategy, and perform a multi-axis synchronous conversion operation to generate a coordinated control instruction according to the control response period.

[0087] The above-mentioned multi-axis coordinated control system 200 of the magnetic suspension motor can implement the multi-axis coordinated control method of the magnetic suspension motor of the above-mentioned method embodiment. The optional items in the above-mentioned method embodiment are also applicable to the embodiment, and will not be described in detail here. The remaining contents of the embodiment of the application can be referred to the contents of the above-mentioned method embodiment, and will not be described in detail in the embodiment.

[0088] The above examples are intended to illustrate and deduce the technical solutions of the present application, and to completely describe the technical solutions, objects and effects of the present application. The purpose is to make the public more thoroughly and comprehensively understand the disclosure of the present application, and does not limit the protection scope of the present application.

[0089] The above examples are not based on an exhaustive enumeration of the present application, and there can be many other unlisted embodiments. Any substitution and improvement made without violating the concept of the present application is within the protection scope of the present application.

Claims

1. A magnetic levitation motor multi-axis coordinated control method, characterized by, The method comprises the following steps: Collecting multi-axis current feedback signals and radial displacement signals of a magnetic suspension motor, performing multi-channel PWM modulation on the multi-axis current feedback signals to obtain PWM modulated current signals, extracting position deviation parameters from the radial displacement signals, coupling and correlating the PWM modulated current signals with the position deviation parameters to generate multi-degree-of-freedom control response values; Using the multi-degree-of-freedom control response values to construct a decoupling control sequence, performing stability detection on the decoupling control sequence to obtain different axial data, performing inter-axis coupling analysis on the different axial data to extract a coupling coefficient sample set, and constructing a control parameter coordination matrix based on the coupling coefficient sample set; Performing control efficiency characteristic analysis on the control parameter coordination matrix to identify an efficient control area, identifying a current load mode based on the multi-axis current feedback signals, selecting a mode matching working point position from the efficient control area according to the current load mode, and determining a coordinated control node by comparing the mode matching working point position with the multi-degree-of-freedom control response values; According to the coordinated control node, control weight distribution correction is performed to determine a conversion parameter set, and a multi-axis coordination sequence is formed by using the conversion parameters through a power switch controller, multi-axis PWM signal synchronization is performed on the multi-axis coordination sequence to generate a multi-axis coordination control strategy; Performing response speed analysis on the multi-axis current feedback signals to obtain dynamic characteristic data, determining a control response period based on the dynamic characteristic data and the multi-axis coordination control strategy, and generating a coordinated control instruction by performing multi-axis synchronous conversion operation according to the control response period.

2. The method of claim 1, wherein, The coupling and correlation of the PWM modulated current signals and the position deviation parameters to generate multi-degree-of-freedom control response values comprises: Performing magnetic field saturation state detection on the PWM modulated current signals to generate saturation state signals and linear state signals; Using the linear state signals to perform nonlinear correction processing on the saturation state signals to form corrected current signals; Performing per-axis control weight distribution on the corrected current signals through the position deviation parameters to generate distributed current vectors; Based on the distributed current vectors, magnetic field saturation compensation control is implemented to obtain multi-degree-of-freedom control response values.

3. The method of claim 1, wherein, The inter-axis coupling analysis on the different axial data to extract a coupling coefficient sample set comprises: Performing control state detection on the different axial data to identify a dominant control axis, an auxiliary control axis and a balance control axis; According to the control coordination relationship between the dominant control axis and the auxiliary control axis, a control signal transmission path is established; With the help of the balance control axis, a buffer adjustment parameter is generated by performing control parameter adjustment and buffering; Based on the buffer adjustment parameter and the control signal transmission path, a coupling coefficient sample set is extracted.

4. The method of claim 1, wherein, The load feature identification based on the multi-axis current feedback signals to determine a current load mode comprises: Using the multi-axis current feedback signals to establish load disturbance detection to generate disturbance characteristic data; Performing disturbance amplitude evaluation on the disturbance characteristic data to find the degree of load mutation; Performing hierarchical control processing on the degree of load mutation to form a buffer control sequence; The current load mode is identified according to the buffer control sequence matching.

5. The method of claim 1, wherein, The mode matching working point position is compared with the multi-freedom degree control response value to determine a coordinated control node, including: The magnetic field loss characteristics of each frequency component in the mode matching working point position are obtained; The multi-freedom degree control response value is matched with the magnetic field loss characteristics to generate a comparison result; The PWM control frequency is optimized based on the comparison result to generate optimal control parameters; The coordinated control strategy is constructed according to the optimal control parameters to determine the coordinated control node.

6. The method of claim 1, wherein, The control weight distribution correction is performed according to the coordinated control node to determine a conversion parameter set, including: The control demand distribution is generated by evaluating the control demand level of each axis for the coordinated control node; The controller is divided into a high-response control area and an energy-saving control area based on the control demand distribution; The controller in the energy-saving control area is switched to a light-load working mode to obtain an energy-saving control signal; The conversion parameter set is determined according to the energy-saving control signal and the high-response control area.

7. The method of claim 1, wherein, The dynamic characteristic data is obtained by analyzing the response speed of the multi-axis current feedback signal, including: The response time component is obtained by detecting the current response time based on the multi-axis current feedback signal; The dynamic response parameters are established by analyzing the dynamic response characteristics of the response time component; The dynamic compensation control signal is generated by performing dynamic control compensation processing according to the dynamic response parameters; The dynamic characteristic data is determined by the dynamic compensation control signal.

8. The method of claim 2, wherein, The saturation state signal and the linear state signal are generated by detecting the magnetic field saturation state of the PWM modulation current signal, including: The current distortion characteristics affected by the switching dead zone are obtained based on the PWM modulation current signal; The actual dead zone time parameters are determined by performing dead zone time backstepping calculation on the current distortion characteristics; The dead zone compensation signal is formed by generating a compensation current according to the actual dead zone time parameters; The saturation state signal and the linear state signal are generated by performing control compensation superposition on the dead zone compensation signal and the PWM modulation current signal.

9. The method of claim 4, wherein, The disturbance characteristic data is generated by establishing load disturbance detection using the multi-axis current feedback signal, including: The change evolution data is obtained by tracking the cross-cycle change for the multi-axis current feedback signal; The disturbance quantization index is established by analyzing the disturbance response amount of each axial load based on the change evolution data, including radial load, axial load and torque load; The disturbance characteristic data is generated by mapping the feature distribution of the disturbance quantization index.

10. A magnetic levitation motor multi-axis coordinated control system, characterized by, The signal acquisition module is configured to acquire multi-axis current feedback signals and radial displacement signals of a magnetic levitation motor, modulate the multi-axis current feedback signals through multiple channels to obtain PWM modulation current signals, extract position deviation parameters from the radial displacement signals, and couple and associate the PWM modulation current signals with the position deviation parameters to generate multi-freedom degree control response values. ​ The coupling analysis module is configured to construct a decoupling control sequence by using the multi-degree-of-freedom control response value, perform stability detection on the decoupling control sequence to obtain different axial data, perform inter-axis coupling analysis by using the different axial data to extract a coupling coefficient sample set, and construct a control parameter coordination matrix based on the coupling coefficient sample set. The load identification module is configured to perform control efficiency feature analysis on the control parameter coordination matrix to identify a high-efficiency control area, perform load feature identification based on the multi-axis current feedback signal to determine a current load mode, select a mode matching working point position from the high-efficiency control area according to the current load mode, compare the mode matching working point position with the multi-degree-of-freedom control response value to determine a coordination control node, and perform control weight distribution correction to determine a conversion parameter set according to the coordination control node. The control coordination module is configured to generate a multi-axis coordination control strategy by using the conversion parameter to compose a multi-axis coordination sequence through a power switch controller, perform multi-axis PWM signal synchronization on the multi-axis coordination sequence, and perform multi-axis synchronous conversion operation according to a control response period determined based on the multi-axis coordination control strategy and dynamic characteristic data obtained by performing response speed analysis on the multi-axis current feedback signal to generate a coordination control instruction. The dynamic adjustment module is configured to perform response speed analysis on the multi-axis current feedback signal to obtain dynamic characteristic data, determine a control response period based on the dynamic characteristic data and the multi-axis coordination control strategy, and perform multi-axis synchronous conversion operation according to the control response period to generate a coordination control instruction.

Images