Magnetic suspension bearing suspension force and zero sequence current decoupling control system

Abstract

The invention discloses a magnetic suspension bearing suspension force and zero sequence current decoupling control system, particularly relates to the technical field of magnetic suspension, and comprises a four-bridge-arm inverter module, a double closed-loop control unit, a multi-source sensing acquisition module, a power driving module and a magnetic suspension bearing body. According to the decoupling control system for the suspension force and the zero-sequence current of the magnetic suspension bearing, the fourth bridge arm connected with the neutral point of the three-phase winding is additionally arranged, an independent zero-sequence current loop is constructed, and zero-sequence current closed-loop control based on the alpha-beta O coordinate system is matched, so that the suspension force and the zero-sequence current of the magnetic suspension bearing are decoupled. Zero-sequence current generated by three-phase imbalance, parameter asymmetry and magnetic field distortion in a traditional three-phase inverter architecture can be pertinently dredged and suppressed; and the fourth bridge arm adopts a SiCMOSFET power switch device and fly-wheel diode combined design, so that a compensation instruction can be quickly responded, winding current distortion caused by superposition of zero-sequence current to working current is avoided, and a linear control relationship between suspension force and current is guaranteed.

Description

A decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing Technical Field

[0001] This invention relates to the field of magnetic levitation technology, and in particular to a decoupling control system for the levitation force and zero-sequence current of a magnetic levitation bearing. Background Technology

[0002] Magnetic levitation bearings, with their advantages of no mechanical contact, low friction loss, and adaptability to high speeds, have been widely used in equipment such as high-speed motors, centrifugal compressors, and magnetic levitation blowers. Their core requirement is precise control of the levitation force to stabilize the rotor position and suppress interference to ensure reliable operation. Currently, most mainstream magnetic levitation bearing control systems adopt a three-phase inverter architecture, but practical applications have revealed many technical bottlenecks that make it difficult to meet the requirements of high precision and high reliability. In traditional three-phase inverters, the winding neutral point is often in a suspended state, which easily generates zero-sequence current due to three-phase current imbalance, winding parameter asymmetry, and air gap magnetic field distortion. This current cannot be released through the three-phase bridge arms and can only be superimposed on the operating current, leading to winding current distortion. This disrupts the linear control relationship between levitation force and current, causing increased rotor displacement deviation, intensified vibration, and even bearing instability.

[0003] Existing control systems are mostly single closed-loop architectures, focusing only on levitation force adjustment and without designing suppression schemes for zero-sequence current. Since there is a strong nonlinear coupling between levitation force and zero-sequence current, it is easy to lose sight of one when adjusting parameters. Focusing on levitation accuracy will lead to zero-sequence current runaway, while suppressing zero-sequence current will sacrifice displacement response performance, making it impossible to achieve dual-objective collaborative optimization.

[0004] The sensor signals such as displacement, current, and magnetic field that the system relies on are susceptible to electromagnetic interference and temperature drift, and lack efficient noise suppression and data fusion mechanisms, which leads to distortion of the state information obtained by the control unit and further reduces the control effect.

[0005] Traditional single-processor architectures need to simultaneously handle tasks such as coupling identification, real-time control, and waveform generation, resulting in high computational load and high control latency. Furthermore, the lack of effective isolation between the control unit and the power drive module makes it easy for strong electrical interference to intrude into weak electrical circuits, causing signal distortion and device damage.

[0006] Weak protection and maintenance functions: Most existing systems only have overcurrent protection and lack comprehensive protection against overvoltage, high temperature and other factors; the communication rate is low and the coverage is limited, making it difficult to achieve remote monitoring, parameter configuration and fault diagnosis, which is not conducive to intelligent operation and maintenance.

[0007] In summary, existing systems have significant shortcomings in zero-sequence current suppression, dual-target decoupling, precise sensing, anti-interference, and comprehensive protection and maintenance, limiting their application in high-precision, high-speed scenarios. Therefore, developing solutions that can effectively decouple levitation force from zero-sequence current, improve control and sensing performance, and enhance system reliability has become an urgent problem to be solved in this field. Summary of the Invention

[0008] The main objective of this invention is to provide a decoupling control system for the levitation force and zero-sequence current of a magnetic levitation bearing, which can effectively solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing, comprising: a four-arm inverter module, a dual-closed-loop control unit, a multi-source sensor acquisition module, a power drive module, and a magnetic levitation bearing body; the four-arm inverter module includes a first three arms and a fourth arm, the first three arms being connected to the levitation force control winding of the magnetic levitation bearing body, and the fourth arm being connected to the neutral point of the three-phase winding of the magnetic levitation bearing body; the multi-source sensor acquisition module is used to acquire the displacement signal, winding current signal, and air gap magnetic field signal of the magnetic levitation bearing, and transmit the acquired signals to the dual-closed-loop control unit; the dual-closed-loop control unit constructs a dual-closed-loop control architecture for levitation force and zero-sequence current based on the αβ0 coordinate system, realizing α-axis levitation force closed-loop control through the first three arms, realizing zero-sequence current closed-loop adjustment through the fourth arm, and synchronously outputting control signals to the power drive module; after receiving the control signals, the power drive module drives the four-arm inverter module to complete the decoupling control of levitation force adjustment and zero-sequence current suppression.

[0010] Preferably, the fourth bridge arm consists of a power switching device and a freewheeling diode connected in series. The power switching device is a SiCMOSFET. The fourth bridge arm forms an independent current loop with the neutral point of the three-phase winding of the magnetic levitation bearing body through an independent neutral point lead-out terminal.

[0011] Preferably, the dual closed-loop control unit includes a nonlinear coupling identification module and a dual-objective control module. The nonlinear coupling identification module adopts an improved ELM extreme learning machine and constructs a nonlinear coupling model of levitation force and zero-sequence current based on the displacement signal, winding current signal and air gap magnetic field signal output by the multi-source sensor acquisition module. The dual-objective control module adopts the Terminal sliding mode control algorithm and takes the levitation displacement tracking accuracy and zero-sequence current suppression effect as dual control objectives to generate levitation force control commands for the first three bridge arms and zero-sequence current compensation commands for the fourth bridge arm.

[0012] Preferably, the multi-source sensing acquisition module includes an eddy current displacement sensor, a Rogowski coil current sensor, a Hall magnetic field sensor, and an NTC temperature sensor. The eddy current displacement sensor is used to acquire the x-axis, y-axis, and z-axis displacement signals of the magnetic levitation bearing rotor. The Rogowski coil current sensor is used to acquire the phase current and zero-sequence current signals of the first three bridge arms. The Hall magnetic field sensor is used to acquire the air gap magnetic field strength signal of the magnetic levitation bearing. The NTC temperature sensor is used to acquire the temperature signal of the levitation force control winding.

[0013] Preferably, an adaptive Kalman filter module is provided between the multi-source sensor acquisition module and the dual closed-loop control unit. This module performs noise suppression and data fusion on the raw signal output by the multi-source sensor acquisition module, and outputs an accurate state perception signal to the dual closed-loop control unit.

[0014] Preferably, the dual closed-loop control unit adopts a CPU+FPGA+MCU heterogeneous multi-core architecture. The CPU is used to run the nonlinear coupling identification module and the adaptive weight adjustment algorithm. The FPGA is used to implement the core operation of dual closed-loop control, PWM waveform generation and high-speed data processing. The MCU is used for sensor data acquisition, operating condition monitoring and fault diagnosis. The CPU and FPGA communicate through the AXI4 bus, and the FPGA and MCU communicate through the SPI interface.

[0015] Preferably, a high-speed optocoupler isolation module is provided between the dual closed-loop control unit and the power drive module to isolate the strong electrical interference signal of the power drive module and protect the weak electrical circuit of the dual closed-loop control unit.

[0016] Preferably, the dual closed-loop control unit is connected to a communication unit, and the power drive module is equipped with a protection unit. The communication unit includes a CPU communication module and a 5G communication module. The CPU communication module is used to configure system parameters and provide operating status feedback to the host computer. The 5G communication module is used to realize high-speed data transmission and remote monitoring. The protection unit includes an overcurrent protection subunit, an overvoltage protection subunit, and a high-temperature protection subunit. The protection unit is linked with the dual closed-loop control unit and adjusts the output current of the four-arm inverter module in real time when an abnormal operating condition is detected.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention constructs an independent zero-sequence current loop by adding a fourth bridge arm connected to the neutral point of the three-phase winding. With the zero-sequence current closed-loop control based on the αβ0 coordinate system, it can specifically guide and suppress the zero-sequence current generated by three-phase imbalance, parameter asymmetry and magnetic field distortion in the traditional three-phase inverter architecture. Moreover, the fourth bridge arm adopts a combination design of SiCMOSFET power switching device and freewheeling diode, which can quickly respond to compensation commands, avoid the winding current distortion caused by the superposition of zero-sequence current to the working current, ensure the linear control relationship between levitation force and current, and reduce rotor displacement deviation and vibration amplitude.

[0018] 2. This invention innovatively adopts a dual closed-loop control architecture of levitation force and zero-sequence current. By constructing a nonlinear coupled model through an improved ELM extreme learning machine, the strong coupling relationship between levitation force and zero-sequence current is accurately identified. Combined with the Terminal sliding mode control algorithm, targeted control commands are generated, and the first three bridge arms focus on levitation force adjustment to ensure displacement tracking accuracy. The fourth bridge arm independently achieves zero-sequence current suppression. This completely changes the shortcomings of the traditional single closed-loop architecture, which suffers from the disadvantage of neglecting one aspect. It can stably control the rotor's x / y / z displacement deviation within the target range and suppress the zero-sequence current to a preset threshold, achieving synergistic optimization of the two control objectives and meeting the application requirements of high-precision and high-speed scenarios.

[0019] 3. The multi-source sensing acquisition module of this invention integrates four types of sensors: displacement, current, magnetic field, and temperature. It can comprehensively capture key state signals during the operation of the magnetic levitation bearing. At the same time, the adaptive Kalman filter module performs noise suppression and data fusion on the original signal. It can not only effectively resist the influence of electromagnetic interference and temperature drift on the sensing signal, but also achieve accurate fusion of multi-dimensional data through weight allocation and timestamp alignment, and output state perception signals with extremely low distortion.

[0020] 4. This invention adopts a heterogeneous multi-core architecture of CPU+FPGA+MCU, which realizes parallel processing of tasks such as coupling identification, core control, and data acquisition through functional division of labor, greatly reducing the computational load of a single processor, shortening control latency, and ensuring real-time response of dual closed-loop control; at the same time, a high-speed optocoupler isolation module is set between the dual closed-loop control unit and the power drive module, which can effectively isolate electromagnetic interference and voltage spikes in the high-voltage circuit, avoid signal distortion and device damage in the low-voltage circuit, and improve the long-term operational reliability of the system.

[0021] 5. The protection unit of this invention integrates triple protection against overcurrent, overvoltage, and high temperature. It can monitor abnormal operating conditions in real time and adjust the output in conjunction with the dual closed-loop control unit to prevent the fault from escalating. The communication unit realizes local parameter configuration and status feedback through the CPU communication module, and achieves high-speed remote data transmission and monitoring by combining with the 5G communication module. This solves the problems of single protection, low communication rate, and inconvenient operation and maintenance in existing systems. It can not only effectively protect equipment safety, but also support remote fault diagnosis and parameter optimization, providing convenient conditions for intelligent operation and maintenance of magnetic levitation bearing equipment and broadening the application scenarios of the system. Attached Figure Description

[0022] Figure 1 is a flowchart of the overall modules and a schematic diagram of the working principle of the present invention. Detailed Implementation

[0023] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0024] As shown in Figure 1, a decoupling control system for the levitation force and zero-sequence current of a magnetic levitation bearing includes: a four-arm inverter module, a dual closed-loop control unit, a multi-source sensor acquisition module, a power drive module, and a magnetic levitation bearing body. The first three arms of the four-arm inverter module are connected one-to-one with the A, B, and C phase levitation force control windings of the magnetic levitation bearing body. Each phase winding achieves current switching and amplitude regulation through the power switching devices of the first three arms. The two power switching devices of the fourth arm are respectively connected to the positive and negative terminals of the DC bus, and their output terminals are led out through independent wires and firmly connected to the common neutral point of the A, B, and C phase windings of the magnetic levitation bearing body, forming a current loop independent of the first three arms. For example, this connection method is used in magnetic levitation high-speed motors. This ensures that the zero-sequence current flows only through the fourth bridge arm without affecting the levitation control current of the first three bridge arms. During the operation of the magnetic levitation bearing, the multi-source sensing module uses an eddy current displacement sensor to output a set of rotor x, y, and z-axis position offset data every 100μs at a sampling frequency of 10kHz. A Rogowski coil current sensor is connected in series in the output lines of the first three bridge arms and the neutral point line of the fourth bridge arm, detecting the instantaneous values ​​of each phase current and the zero-sequence current with an accuracy of 0.1A within a measurement range of -50A to +50A. A Hall effect magnetic field sensor is installed at three evenly distributed sampling points in the air gap, detecting the magnetic induction intensity with a resolution of 0.001T within a range of 0.1T to 1T. All acquired signals are aligned by timestamps and transmitted at a high-speed digital transmission rate of 1Mbps. The data is serially transmitted to the dual-closed-loop control unit via the bus, with a transmission delay of no more than 1ms. The dual-closed-loop control unit first converts the acquired three-phase current and displacement signals to the αβ0 coordinate system (α axis coincides with the axis of phase A winding, β axis is perpendicular to α axis, and 0 axis is the zero-sequence component channel). In this coordinate system, a levitation force control closed loop and a zero-sequence current control closed loop are constructed. In the levitation force control closed loop, the target displacement deviations of the 5kW magnetic levitation motor in the x and y directions (±0.01mm) and z direction (±0.02mm) are calculated with the actual displacement signals. If the absolute value of the deviation is greater than 0.005mm, the winding current amplitude is changed by adjusting the duty cycle of the power switching devices in the first three bridge arms to bring the deviation closer to zero. In the zero-sequence current control closed loop, the actual value of the zero-sequence current is compared with the ±0.01mm target displacement deviations in the x and y directions (±0.01mm and ±0.02mm). The difference is calculated based on the target suppression value of 0.5A. The conduction time of the power switching device of the fourth bridge arm is adjusted according to the difference. For example, when the zero-sequence current is 1A, the duty cycle of the lower bridge arm is increased to generate a reverse compensation current. The two closed loops synchronously generate PWM control signals with a 100μs operation cycle and transmit them to the power drive module. After receiving the PWM control signal output by the dual closed-loop control unit, the power drive module first amplifies the 5V weak current control signal into a 15V, 1A drive signal. Then, according to the duty cycle command, it controls the power switching devices of the first three bridge arms to alternately turn on and off at a frequency of 20kHz to adjust the current of each phase winding to achieve precise adjustment of the levitation force. At the same time, it controls the fourth bridge arm to act according to the zero-sequence current compensation command. For example, the load change of the magnetic levitation blower causes the zero-sequence current to rise from 0.3A to 0.At 8A, the fourth bridge arm suppresses it to 0.4A within 2ms, ultimately driving the four-bridge arm inverter module to complete the decoupled control of levitation force regulation and zero-sequence current suppression.

[0025] Furthermore, in the above description, the fourth bridge arm consists of a series-connected power switching device and a freewheeling diode. In the circuit structure of the fourth bridge arm, two SiCMOSFET power switching devices are connected in series as upper and lower bridge arms. The drain of the upper bridge arm SiCMOSFET is connected to the positive terminal of the DC bus, and its source is connected to the drain of the lower bridge arm SiCMOSFET. The source of the lower bridge arm SiCMOSFET is connected to the negative terminal of the DC bus. A freewheeling diode is connected in reverse parallel between the source and drain of each SiCMOSFET power switching device. The anode of the freewheeling diode is connected to the source of the SiCMOSFET, and the cathode is connected to the drain of the SiCMOSFET. During the start-up and shutdown process of the magnetic levitation bearing, when the SiCMOSFET is suddenly turned off, the reverse electromotive force generated by the winding inductance can be released through the freewheeling diode, allowing the current to decay smoothly and preventing voltage spikes from damaging the device. The power switching device uses SiCMOSFET... The fourth bridge arm of the SFET forms an independent current loop with the neutral point of the three-phase winding of the magnetic levitation bearing body through an independent neutral point lead-out terminal. The independent neutral point lead-out terminal of the fourth bridge arm is a copper wire with a cross-sectional area of ​​2mm². One end is soldered to the connection point of the upper and lower bridge arm SiCMOSFETs, and the other end is fixed to the silver-plated neutral point terminal of the three-phase winding of the magnetic levitation bearing body with bolts. The connection line is arranged separately and kept at a distance of more than 5cm from the winding lines of the first three bridge arms. The outer side is fitted with a shielding tube to reduce electromagnetic interference. In the independent current loop formed, the current flows out from the positive terminal of the DC bus, through the upper bridge arm SiCMOSFET, the independent neutral point lead-out terminal, the neutral point of the three-phase winding, the three-phase winding, and the power switching devices of the first three bridge arms, or through the lower bridge arm SiCMOSFET of the fourth bridge arm and the negative terminal of the DC bus to form a loop, ensuring that the zero-sequence current only flows through the fourth bridge arm and does not affect the levitation force control loop of the first three bridge arms.

[0026] Furthermore, the dual closed-loop control unit includes a nonlinear coupling identification module and a dual-target control module. During operation of the magnetically levitated centrifugal compressor, the nonlinear coupling identification module receives the displacement signal, winding current signal, and air gap magnetic field signal after adaptive Kalman filtering. It first processes the displacement signal (from -0.1mm to 0.1mm) according to: Mapped to [-1,1], current signal (-20A to 20A), press: Mapped to [-1,1], magnetic field signal (0.2T to 0.8T), by: Mapped to [0,1], 10,000 normalized signals from the past 1 second are selected as samples (8,000 training sets and 2,000 test sets). Each sample contains 6 input parameters (x, y, z displacement deviations, A-phase current, B-phase current, and magnetic induction intensity) and 2 output parameters (levitation force and zero-sequence current). The intermediate output matrix is ​​obtained by calculating the mapping relationship between the input samples of the training set and the hidden layer nodes. Then, the model is constructed by solving the output weight matrix through Moore-Penrose generalized inverse. For example, when the input x-axis displacement deviation is 0.003mm, A-phase current is 10A, and magnetic induction intensity is 0.5T, the model outputs a levitation force of 500N and a zero-sequence current of 0.3A. The absolute value of the prediction error is controlled within 5%. When the error exceeds the limit, the parameters are updated by reselecting the samples from the most recent 0.5s. The dual-objective control module takes the levitation displacement tracking accuracy and the zero-sequence current suppression effect as dual control objectives and constructs a levitation displacement sliding surface function. ;according to The derivative sign adjusts the duty cycle of the first three bridge arms. A positive derivative decreases the duty cycle, while a negative derivative increases it. The adjustment amount is: , ( s, when exceeding the standard, increases to s), with the duty cycle limited to 1096-90%; simultaneously, a zero-sequence current sliding mode surface function is constructed: ( For zero-sequence current deviation, c2=3, d=1, e=3) when At time A, ,according to Derivative sign and Adjust the duty cycle of the fourth bridge arm with positive and negative values; adjustment amount: d d ( s, when exceeding the standard, increases to The duty cycle is limited to 5%-9596, and finally the suspension force control command of the first three bridge arms and the zero-sequence current compensation command of the fourth bridge arm are generated.

[0027] Furthermore, the multi-source sensing acquisition module includes eddy current displacement sensors, Rogowski coil current sensors, Hall effect magnetic field sensors, and NTC temperature sensors. There are three eddy current displacement sensors, respectively positioned at the stator's x, y, and z axes detection positions. The initial air gap between the probe and the rotor surface is 2mm. The output voltage is 0-10V corresponding to a displacement of -0.1mm to 0.1mm. When the rotor's x-axis displacement is 0.003mm, the output is 5.15V. The signal is transmitted via a shielded cable to a conditioning circuit for filtering and amplification before being sent to the dual closed-loop control unit. The eddy current displacement sensors are used to acquire data on the x, y, and z axes of the magnetic levitation bearing rotor. The z-axis displacement signal is generated by four Rogowski coil current sensors: three are mounted on the A, B, and C phase cables of the first three bridge arms, and one is mounted on the neutral point cable of the fourth bridge arm. Each coil has 1000 turns and a 1000Ω sampling resistor. The induced voltage is converted to a DC voltage signal using a 1V / A conversion factor. When the A-phase current is 12.5A, the output is 12.5V, allowing real-time monitoring of the current in each phase. The Rogowski coil current sensors are used to collect the phase currents and zero-sequence current signals of the first three bridge arms. Hall effect magnetic field sensors are installed at three evenly distributed positions in the stator slots, outputting a 0-5V voltage corresponding to a magnetic induction intensity of 0.2T to 0.8T. T T, conversion, 3V output corresponds to 0.56T, reflecting the magnetic field distribution; Hall effect magnetic field sensor is used to collect the magnetic field strength signal in the air gap of magnetic levitation bearing; NTC temperature sensor is patch-mounted at the ends of the A, B, and C phase windings, with a resistance of 10kΩ at 25℃ and a temperature coefficient of -3% / ℃. The temperature measurement resistance is calculated to be 8.5kΩ, corresponding to a measurement range of -40℃ to 125℃ and an accuracy of ±1℃ for monitoring winding temperature. The NTC temperature sensor is used to collect the levitation force control winding temperature signal.

[0028] Furthermore, an adaptive Kalman filter module is set between the multi-source sensor acquisition module and the dual closed-loop control unit. This module receives four types of discrete time series signals output by the multi-source sensor acquisition module and establishes the state equation: (Where A is the identity matrix, B is the zero matrix, and W is Gaussian white noise with a variance of 0.01) and the observation equation: (Where H is the identity matrix and V is Gaussian white noise with a variance of 0.02), calculate the prior state estimate at time k: And the prior error covariance matrix: Then calculate the Kalman gain: Update the posterior state estimate: And the posterior error covariance matrix: The observation residuals were calculated using the sliding window method (100 sampling points). variance , Then multiply R and Q by 0.8 and 0.9 respectively. The time remains unchanged. The values ​​are multiplied by 1.2 and 1.1 respectively. For example, a displacement signal noise amplitude of 0.01mm is reduced to ≤ 0.001mm after filtering. During data fusion, similar signals are treated as follows: Calculate the weights (where σ is the standard deviation of the error). When the temperature sensor in phase A has σ = 0.5℃, phase B has σ = 0.6℃, and phase C has σ = 0.5℃, the weights are 0.37, 0.26, and 0.37 respectively. Fusion temperature: Different types of signals are aligned according to their timestamps to construct state vectors: The data is transmitted in parallel to the dual closed-loop control unit within 50μs. The raw signal output by the multi-source sensor acquisition module is subjected to noise suppression and data fusion, and a precise state perception signal is output to the dual closed-loop control unit.

[0029] Furthermore, the dual-loop control unit adopts a heterogeneous multi-core architecture of CPU+FPGA+MCU. The CPU uses an ARM Cortex-A9 dual-core processor. The nonlinear coupling identification module program is stored in Flash and loaded into memory during runtime. It reads the filtered data transmitted by the FPGA at 50ms cycles, performs model parameter update calculations, and runs an adaptive weight adjustment algorithm. When the suspension displacement tracking error is greater than 0.008mm for 5 consecutive cycles, the displacement signal weight is adjusted from 0.3 to 0.4, and the magnetic field signal weight is adjusted from 0.2 to 0.1. The CPU is used to run the nonlinear coupling identification module and the adaptive weight adjustment algorithm. The FPGA uses a Xilinx Artix-7 chip. The dual-loop control core operation takes ≤1μs per step. The PWM waveform generation module generates a 20kHz PWM signal with a resolution of 0.1% according to the duty cycle instruction. The high-speed data processing module receives the raw signal at 100μs cycles and transmits it to the filtering module. At the same time, it transmits the accurate signal to the CPU and MCU at a transmission rate of 100Mbps. The MCU (Application Processor) is used to implement the core operation of dual closed-loop control, PWM waveform generation, and high-speed data processing. It uses an STM32F103 microcontroller, which acquires sensor signals via a 12-bit ADC interface, transmits them to RAM via DMA, and verifies their validity. The operating condition monitoring module reads data in real time, identifying abnormalities when the winding temperature exceeds 85℃, the current exceeds 30A, or the displacement deviation exceeds 0.02mm. The fault diagnosis module identifies faults such as sensor signal loss and sudden current changes and sends signals to the FPGA. The MCU is used for sensor data acquisition, operating condition monitoring, and fault diagnosis. The CPU and FPGA communicate via an AXI4 bus; the CPU transmits model parameters and control commands to the FPGA, and the FPGA transmits filtered data and control error data to the CPU. The CPU and FPGA communicate via an AXI4 bus, and the FPGA and MCU communicate via an SPI interface. The FPGA transmits target parameters and control status signals to the MCU, and the MCU transmits raw sensor data and fault diagnosis results to the FPGA. The communication cycle is 100μs. The FPGA and MCU communicate via an SPI interface.

[0030] Furthermore, a high-speed optocoupler isolation module is set between the dual closed-loop control unit and the power drive module, using a 6N137 high-speed optocoupler. The 5V, 20mA low-voltage control signal output from the dual closed-loop control unit is input to the optocoupler input terminal via a 220Ω current-limiting resistor. When the signal is high, the input current is approximately 17.3mA, causing the internal LED to light up. The optocoupler output terminal adopts an open-collector design, with an external 10kΩ pull-up resistor connected to the 15V drive power supply. When there is current at the input terminal, the output is low; when there is no current, the output is high, realizing the transition from low voltage to high voltage. The electrical signal isolation and conversion, for example, the optocoupler outputs 0.2V when the control signal is high, triggering the driver chip to output a drive signal; the module isolation voltage is ≥2500Vrms, which can block electromagnetic interference and voltage spikes generated by the 500V high-voltage circuit of the power drive module. For example, after isolation, the induced voltage of a 200V voltage spike in the low-voltage circuit is ≤0.1V, and the transmission delay is ≤50ns, meeting the requirements for 20kHz PWM signal transmission. It is used to isolate the high-voltage interference signal of the power drive module and protect the low-voltage circuit of the dual closed-loop control unit.

[0031] Furthermore, the dual closed-loop control unit is connected to a communication unit, and the power drive module is equipped with a protection unit. The communication unit includes a CPU communication module and a 5G communication module. The CPU communication module is connected to an industrial switch via an Ethernet interface, establishing a TCP / IP connection with the host computer. The host computer sends " =0.00mm, Configuration commands such as "=25A,k1=0.0196 / s" are executed. The CPU verifies the CRC32 checksum, parses and stores the parameters, reads the register to verify consistency, and then sends a "configuration successful" message. Simultaneously, it sends JSON-formatted status data (e.g., JSON data) to the host computer every 1 second. “ "0.002mm," "12.3A", "0.2A, "T": 45°C, "D1" 45.296}); The CPU communication module is used to configure system parameters and provide operational status feedback to the host computer. The 5G communication module adopts an industrial-grade dual-mode module, which connects to the CPU via a PCIe interface. The data transmission rate is >10Mbps, transmitting 1KB data packets to the remote platform every 500ms. It supports remote start / stop and parameter adjustment command reception, with a response time of The 5G communication module is used to achieve high-speed data transmission and remote monitoring. The protection unit includes an overcurrent protection subunit, an overvoltage protection subunit, and a high-temperature protection subunit. The overcurrent protection subunit samples the current through a 0.01Ω sampling resistor, and the comparator outputs a trigger signal when the voltage drop reaches 20.3V. The overvoltage protection subunit samples the bus voltage through a 1:100 voltage divider resistor, and triggers when the divided voltage is ≥5.5V (corresponding to 550V). The high-temperature protection subunit samples the heatsink temperature through an NTC sensor, and triggers when it is ≥85℃. The protection unit is linked with the dual closed-loop control unit. When an abnormal operating condition is detected, the FPGA receives the trigger signal and, in case of overcurrent, presses the following: Adjust the duty cycle every 1ms until the current is 10A; during overvoltage, reduce the duty cycle of all bridge arms to 2096, and if it does not recover, reduce it by another 10% until the voltage is <520V; during high temperature, press... Adjust the output current of the four-arm inverter module in real time, adjusting it every 500ms until the temperature is ≤80℃.

[0032] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A decoupling control system for levitation force and zero-sequence current in a magnetic levitation bearing, characterized in that, include: The system comprises a four-arm inverter module, a dual-loop control unit, a multi-source sensor acquisition module, a power drive module, and a magnetic levitation bearing body. The four-arm inverter module includes three front arms and a fourth arm. The three front arms are connected to the levitation force control winding of the magnetic levitation bearing body, and the fourth arm is connected to the neutral point of the three-phase winding of the magnetic levitation bearing body. The multi-source sensor acquisition module acquires the displacement signal, winding current signal, and air gap magnetic field signal of the magnetic levitation bearing and transmits the acquired signals to the dual-loop control unit. The dual-loop control unit constructs a dual-loop control architecture for levitation force and zero-sequence current based on the αβ0 coordinate system. It achieves α-axis levitation force closed-loop control through the three front arms and zero-sequence current closed-loop regulation through the fourth arm, synchronously outputting control signals to the power drive module. Upon receiving the control signals, the power drive module drives the four-arm inverter module to complete the decoupled control of levitation force regulation and zero-sequence current suppression.

2. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 1, characterized in that: The fourth bridge arm consists of a power switching device and a freewheeling diode connected in series. The power switching device is a SiCMOSFET. The fourth bridge arm forms an independent current loop with the neutral point of the three-phase winding of the magnetic levitation bearing body through an independent neutral point lead-out terminal.

3. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 1, characterized in that: The dual closed-loop control unit includes a nonlinear coupling identification module and a dual-objective control module. The nonlinear coupling identification module adopts an improved ELM extreme learning machine and constructs a nonlinear coupling model of levitation force and zero-sequence current based on the displacement signal, winding current signal and air gap magnetic field signal output by the multi-source sensor acquisition module. The dual-objective control module adopts the Terminal sliding mode control algorithm and takes the levitation displacement tracking accuracy and zero-sequence current suppression effect as dual control objectives to generate levitation force control commands for the first three bridge arms and zero-sequence current compensation commands for the fourth bridge arm.

4. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 1, characterized in that: The multi-source sensing acquisition module includes an eddy current displacement sensor, a Rogowski coil current sensor, a Hall magnetic field sensor, and an NTC temperature sensor. The eddy current displacement sensor is used to acquire the x, y, and z displacement signals of the magnetic levitation bearing rotor. The Rogowski coil current sensor is used to acquire the phase current and zero-sequence current signals of the first three bridge arms. The Hall magnetic field sensor is used to acquire the magnetic field strength signal of the air gap of the magnetic levitation bearing. The NTC temperature sensor is used to acquire the temperature signal of the levitation force control winding.

5. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 4, characterized in that: An adaptive Kalman filter module is set between the multi-source sensor acquisition module and the dual closed-loop control unit. This module performs noise suppression and data fusion on the raw signal output by the multi-source sensor acquisition module, and outputs an accurate state perception signal to the dual closed-loop control unit.

6. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 3, characterized in that: The dual-closed-loop control unit adopts a heterogeneous multi-core architecture of CPU+FPGA+MCU. The CPU is used to run the nonlinear coupling identification module and the adaptive weight adjustment algorithm. The FPGA is used to implement the core operation of dual-closed-loop control, PWM waveform generation and high-speed data processing. The MCU is used for sensor data acquisition, operating condition monitoring and fault diagnosis. The CPU and FPGA communicate through the AXI4 bus, and the FPGA and MCU communicate through the SPI interface.

7. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 1, characterized in that: A high-speed optocoupler isolation module is provided between the dual closed-loop control unit and the power drive module to isolate the strong electrical interference signal of the power drive module and protect the weak electrical circuit of the dual closed-loop control unit.

8. The decoupling control system for levitation force and zero-sequence current of a magnetic levitation bearing according to claim 1, characterized in that: The dual closed-loop control unit is connected to a communication unit, and the power drive module is equipped with a protection unit. The communication unit includes a CPU communication module and a 5G communication module. The CPU communication module is used to configure system parameters and provide operating status feedback to the host computer. The 5G communication module is used to realize high-speed data transmission and remote monitoring. The protection unit includes an overcurrent protection subunit, an overvoltage protection subunit, and a high-temperature protection subunit. The protection unit is linked with the dual closed-loop control unit. When an abnormal operating condition is detected, the output current of the four-bridge inverter module is adjusted in real time.

Images