Intelligent start-stop control method and system for magnetic suspension refrigeration compressor

Abstract

The invention discloses an intelligent start-stop control method and system for a magnetic suspension refrigeration compressor, and the method comprises the steps: obtaining a position perception enhancement source through collecting and analyzing the current response data of a magnetic bearing coil, and extracting the current response change characteristics; rotor position estimation is carried out based on current response change characteristics, a positioning precision driving source is extracted from estimation deviation, and a non-impact starting excitation current instruction is generated; a multi-sensor fusion data network is established, configuration is reconstructed through an abnormal response activation signal, and a degradation operation fault-tolerant mode is formed; converting a response preparation gain according to the shutdown instruction signal, performing shutdown state evaluation in combination with signal reconstruction configuration, and generating an emergency shutdown trigger condition; and the power supply state of a bus capacitor is detected, gradient control potential is released, the active damping control opportunity is determined, a damping force control instruction is generated, finally intelligent start-stop control over the magnetic suspension refrigeration compressor is achieved, and safe and reliable operation of the system under various working conditions is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of compressor control, in particular to an intelligent start-stop control method and system of a magnetic suspension refrigeration compressor. BACKGROUND

[0002] The magnetic suspension refrigeration compressor adopts electromagnetic bearings to support the rotor, realizes non-contact suspension operation, has the advantages of no friction, no need for lubrication, low vibration and noise, and is widely used in the fields of refrigeration and air conditioning, industrial compression, etc. However, the start-stop process of the magnetic suspension compressor is a key link of system operation, and improper start-stop control can cause serious problems such as rotor collision with standby bearings, vibration overrun, position instability, etc., affecting the service life and operation reliability of the equipment.

[0003] The existing start-stop control method of the magnetic suspension compressor mainly relies on preset control parameters and fixed start-stop procedures, and lacks adaptive adjustment capability for real-time operation state. In the starting process, the sudden change of excitation current can easily cause impact and vibration; in the shutdown process, the utilization of bus capacitor energy is not sufficient, and smooth deceleration cannot be achieved; at the same time, in the case of sensor drift, signal anomaly and other fault conditions, the system lacks effective fault tolerance mechanism, and it is difficult to ensure safe and reliable start-stop control. Therefore, it is urgent to develop an intelligent start-stop control method that can realize impact-free start and smooth stop. SUMMARY

[0004] The present application discloses an intelligent start-stop control method and system for a magnetic suspension refrigeration compressor, which aims to obtain a position perception enhancement source through fluctuation analysis of current response data, accurately estimate the rotor position and generate an impact-free start excitation current instruction; establish a multi-sensor fusion data network, form a degraded operation fault tolerance mode through signal reconstruction configuration; convert the shutdown instruction signal into a response preparation gain, combine bus capacitor energy gradient control and active damping control, and realize intelligent start-stop control of the magnetic suspension refrigeration compressor.

[0005] The present application discloses an intelligent start-stop control method and system for a magnetic suspension refrigeration compressor, which aims to obtain a position perception enhancement source through fluctuation analysis of current response data, accurately estimate the rotor position and generate an impact-free start excitation current instruction; establish a multi-sensor fusion data network, form a degraded operation fault tolerance mode through signal reconstruction configuration; convert the shutdown instruction signal into a response preparation gain, combine bus capacitor energy gradient control and active damping control, and realize intelligent start-stop control of the magnetic suspension refrigeration compressor. Collecting current response data of the magnetic bearing coil, performing fluctuation analysis on the current response data to obtain a position perception enhancement source, and obtaining current response change characteristics based on the position perception enhancement source; Performing rotor position estimation on the current response change characteristics to extract rotor offset coordinate parameters, extracting a positioning accuracy driving source from the estimation deviation of the rotor offset coordinate parameters, and generating an impact-free start excitation current instruction based on the positioning accuracy driving source processing the current response change characteristics; A fusion data network is established based on the non-impact starting excitation current instruction to monitor the multi-sensor operating state, an abnormal response activation signal reconstruction configuration is activated in the fusion data network, and a degraded operation fault-tolerant mode is generated based on the signal reconstruction configuration in combination with the current response change characteristics; A shutdown instruction signal is received through the degraded operation fault-tolerant mode, the shutdown instruction signal is converted into a response preparation gain, a shutdown state evaluation is performed in combination with the signal reconstruction configuration according to the response preparation gain to generate an emergency shutdown trigger condition; The bus capacitor power supply state is detected according to the emergency shutdown trigger condition, a gradient control potential is released from the energy change of the bus capacitor power supply state, a proactive damping control opportunity is determined based on the gradient control potential in combination with the degraded operation fault-tolerant mode, a damping force control instruction is generated through the proactive damping control opportunity, and intelligent start-stop control of the magnetic suspension refrigeration compressor is realized.

[0006] The second aspect of the application provides an intelligent start-stop control system of a magnetic suspension refrigeration compressor, comprising: A data acquisition module is configured to acquire current response data of a magnetic bearing coil, perform fluctuation analysis on the current response data to obtain a position perception enhancement source, and obtain current response change characteristics based on the position perception enhancement source. A position estimation module is configured to perform rotor position estimation on the current response change characteristics to extract rotor offset coordinate parameters, extract a positioning accuracy driving source from the estimated deviation of the rotor offset coordinate parameters, and generate a non-impact starting excitation current instruction based on the positioning accuracy driving source processing the current response change characteristics. A fault monitoring module is configured to monitor the multi-sensor operating state based on the non-impact starting excitation current instruction to establish a fusion data network, activate a signal reconstruction configuration through an abnormal response in the fusion data network, and generate a degraded operation fault-tolerant mode based on the signal reconstruction configuration in combination with the current response change characteristics. A shutdown control module is configured to receive a shutdown instruction signal through the degraded operation fault-tolerant mode, convert the shutdown instruction signal into a response preparation gain, perform shutdown state evaluation in combination with the signal reconstruction configuration according to the response preparation gain to generate an emergency shutdown trigger condition. A damping execution module is configured to detect the bus capacitor power supply state according to the emergency shutdown trigger condition, release a gradient control potential from the energy change of the bus capacitor power supply state, determine a proactive damping control opportunity based on the gradient control potential in combination with the degraded operation fault-tolerant mode, generate a damping force control instruction through the proactive damping control opportunity, and realize intelligent start-stop control of the magnetic suspension refrigeration compressor.

[0007] The beneficial effects of this invention are reflected in the following points: First, by constructing a position sensing enhancement source through current response fluctuation analysis, indirect measurement of rotor position based on current information is realized, overcoming the measurement limitations of traditional position sensors in high-speed rotation environments. In this process, the estimated deviation of rotor position is no longer a simple error, but is extracted as a driving source for positioning accuracy. Its change law directly guides the generation of excitation current commands, enabling the control system to adaptively adjust according to actual deviation characteristics, eliminating the start-up impact problem of fixed parameter control. Second, a complete fault detection and compensation mechanism is established by combining a multi-sensor fusion data network with signal reconstruction configuration technology. The identification of sensor drift modes is not only used for fault diagnosis, but also becomes the basis for triggering collaborative calibration. A set of compensation coordinates is generated through multi-channel cross-verification. The degraded operation fault-tolerant mode automatically adjusts the control strategy according to the level of abnormal response, transforming the abnormal state information of the system into the decision basis for fault-tolerant control, ensuring stable operation even when some sensors fail. Finally, the torque standing wave formation technology redistributes the rotor's oscillation energy. By injecting a counter-phase torque at a specific phase, the originally disordered oscillations are transformed into a stable standing wave mode, and the standing wave node becomes the ideal shutdown target location. The energy decay process of the bus capacitor is divided into different control regions through gradient analysis. The remaining energy in each region is fully utilized for the corresponding control task, thereby optimizing the energy distribution during the shutdown process and extending the controllable shutdown time window. Attached Figure Description

[0008] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0009] Unless otherwise specified, the same reference numerals in different figures represent the same or similar technical features, and different reference numerals may be used to represent the same or similar technical features.

[0010] Figure 1 This is a flowchart illustrating an intelligent start-stop control method for a magnetic levitation refrigeration compressor according to the present invention.

[0011] Figure 2 This is a structural block diagram of an intelligent start-stop control system for a magnetic levitation refrigeration compressor according to the present invention. Detailed Implementation

[0012] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0013] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0014] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0015] The technical solutions of the embodiments of this application will be described below.

[0016] like Figure 1 As shown, this embodiment of the invention provides an intelligent start-stop control method for a magnetic levitation refrigeration compressor, including the following steps S110-S150: Step S110: Collect current response data of the magnetic bearing coil, perform fluctuation analysis on the current response data to obtain the position sensing enhancement source, and obtain the current response change characteristics based on the position sensing enhancement source.

[0017] Specifically, the current response data of the magnetic bearing coils is collected. High-precision current sensors are deployed in the control loop of the magnetic bearing system to collect the excitation current signals of the radial and axial magnetic bearing coils in real time. The current sensors are Hall effect type sensors with a measurement range of 0-50A, a resolution of 0.001A, and a response time of less than 1 microsecond, ensuring the ability to capture rapid changes in current. The radial magnetic bearing contains four orthogonally distributed electromagnetic coils, each equipped with an independent current sensor, forming four-path control current signals: X-positive, X-negative, Y-positive, and Y-negative. The axial magnetic bearing contains two sets of thrust coils, upper and lower, which monitor the upper thrust current and lower thrust current respectively. Current data acquisition adopts a synchronous sampling mode. All sensors achieve strict time synchronization through hardware trigger signals, and the sampling frequency is set to 20kHz to ensure that the requirements of the Nyquist sampling theorem are met. The acquisition process continuously records the rotor's current response under different operating conditions, including typical conditions such as the start-up phase, steady-state operation, acceleration process, load change, and shutdown process. Current data is continuously collected for 10 minutes under each operating condition to form a complete current response dataset containing time series, current amplitude, and phase information.

[0018] In some embodiments, the step of performing fluctuation analysis on the current response data to obtain a position awareness enhancement source includes: performing frequency anomaly detection on the current response data to obtain inductor field fluctuation characteristics; identifying a position error compensation current sequence in the inductor field fluctuation characteristics; performing amplitude calibration on the position error compensation current sequence to form rotor positioning characteristics; and constructing a position awareness enhancement source using the rotor positioning characteristics.

[0019] Frequency anomaly detection was performed on the current response data to obtain inductor field fluctuation characteristics. Frequency anomaly detection used the Welch method to estimate the power spectrum of the current signal, with a segment length of 4096 points, an overlap rate of 50%, and a Hanning window as the window function. The baseline for the normal frequency distribution was determined using steady-state operating data, with the main energy concentrated in the 0-2kHz frequency band, and the peak frequencies corresponding to the rotor's rotational frequency and its harmonics. Anomaly frequency components were identified through spectral peak detection; a frequency point with a power density exceeding three times the neighborhood mean was marked as an anomalous frequency peak. Inductor field fluctuation characteristics manifested as abnormal energy concentration within a specific frequency band, typically occurring in the 500-1500Hz range, corresponding to the modulation effect generated by the interaction between the magnetic field and the rotor. Fluctuation feature extraction included parameters such as the center frequency, bandwidth, energy proportion, and occurrence time of the anomalous frequency. The center frequency was determined using a weighted average method, with the weight being the power density value of each frequency point. The bandwidth was determined using the -3dB cutoff frequency, reflecting the spectral width of the anomalous frequency components. The energy percentage is calculated by integrating the ratio of energy in the abnormal frequency band to the total energy; a percentage exceeding 10% indicates significant inductive field fluctuations. Time-domain features are extracted using bandpass filtering, with the filter's center frequency set to the detected abnormal frequency and its bandwidth 1.5 times the bandwidth of the abnormal frequency.

[0020] Identifying position error compensation current sequences from inductor field fluctuation characteristics. The identification of position error compensation currents begins with the 500-1500Hz anomalous frequency band of the fluctuation characteristics, which corresponds to the characteristic range of magnetic field modulation effects. The compensation current sequence exhibits a specific energy distribution pattern within this band, with frequencies where the energy percentage exceeds 10% typically containing the compensation signal. The identification process uses the center frequency of the inductor field fluctuation characteristics as a filter design parameter, with the bandwidth set to the detected anomalous bandwidth value. The radial X-direction compensation current is extracted from the X component of the fluctuation characteristics, and the modulation components within the 100-1000Hz range are separated using bandpass filtering. The radial Y-direction compensation current is obtained from the Y component fluctuations, and its spectral distribution is orthogonal to the X component. The axial compensation current is identified from the Z-direction fluctuation characteristics; its frequency characteristics differ significantly from the radial component, mainly concentrated in the low-frequency band. The polarity of the compensation current is determined by the phase information of the fluctuation characteristics; a positive phase corresponds to a positive compensation force, and a negative phase corresponds to a negative compensation force. The continuity of the compensation current sequence is determined by the time-domain envelope of the fluctuation characteristics; the smoothness of the envelope reflects the stability of the compensation process. The sequence extraction considers a -3dB bandwidth range for fluctuation characteristics to ensure the inclusion of the main compensation energy components. The identification results retain the time stamp of the fluctuation characteristics, forming a compensation current sequence with precise timestamps.

[0021] The rotor positioning characteristics are formed by amplitude calibration of the position error compensation current sequence. The position error compensation current sequence, containing time sequences of compensation current in the X, Y, and Z directions, is used as the calibration input. The amplitude calibration determines the conversion coefficients based on the force-current and force-displacement characteristics of the magnetic bearing. The electromagnetic force generated by the magnetic bearing is F = Ki × i / g, where Ki is the current stiffness coefficient, i is the instantaneous value in the compensation current sequence, and g is the air gap length. The relationship between displacement and current is x = (Ki × i) / (Kx × g), where x is the rotor displacement and Kx is the displacement stiffness. This relationship is used to convert the compensation current sequence point by point into a displacement sequence. The calibration coefficients are determined using a system identification method. The current response is measured under known displacement excitation, and the conversion coefficients are obtained through least-squares fitting. The radial positioning characteristics include the X and Y coordinate time sequences of the rotor center, with an accuracy of ±1 micrometer, obtained by conversion from the X and Y direction compensation current sequences respectively. The axial positioning characteristics include the Z coordinate sequence of the rotor's axial position, with an accuracy of ±2 micrometers, obtained by conversion from the Z direction compensation current sequence. The rotor attitude characteristics include tilt angles α and β, which are obtained by calculating the difference between the displacement sequences at adjacent time points and the measurement span, with an angular resolution of 0.01 milliradians.

[0022] A position-aware enhancement source is constructed using rotor positioning features. The source is built from calibrated X, Y, and Z coordinates and tilt angles α and β, expanding to form a complete six-degree-of-freedom motion description of the rotor. Translational degrees of freedom are directly obtained from calibration using X (±1 μm accuracy), Y (±1 μm accuracy), and Z (±2 μm accuracy), reflecting the real-time spatial position of the rotor's center of mass. Rotational degrees of freedom are obtained through geometric transformations of calibrated tilt angles α and β; the pitch angle corresponds to the sine component of angle α, the yaw angle corresponds to the sine component of angle β, and the roll angle is calculated through the coupling relationship between α and β. Motion trajectory reconstruction connects discrete X and Y coordinate points sequentially over time to form the rotor's whirl trajectory in the radial plane; changes in the Z coordinate form the axial vibration curve. Trajectory features are extracted from the reconstructed position data; the ellipse fitting parameters of the X and Y plane trajectory reflect the whirl amplitude and direction, while the amplitude statistics in the Z direction reflect axial stability. The 0.01 milliradian angular resolution obtained from calibration supports accurate attitude change detection; angular velocity is obtained through the angular difference between adjacent time points. The enhancement process combines these basic positioning parameters with their time derivatives to generate a complete state vector containing position, velocity, and acceleration. The fusion process assigns a higher weight (0.8) to low-frequency position information and a higher weight (0.6) to high-frequency attitude information, forming a multi-scale position awareness enhancement source.

[0023] Current response variation characteristics are acquired based on a position-aware augmentation source. The multi-dimensional feature vectors from the augmentation source are used as the analysis benchmark, and the corresponding response patterns are extracted by comparing them with the original current signal. Amplitude variation characteristics are demodulated using the position fluctuation information of the augmentation source. When a radial displacement of 1 micrometer is detected, the corresponding current change is approximately 0.2A; this correspondence is extracted using the Hilbert envelope. Phase offset characteristics are calculated by referring to the phase of the augmentation source's frequency conversion component; the deviation angle reflects the response delay of the control loop. Frequency modulation characteristics are derived from the 100-1000Hz modulation frequencies identified by the augmentation source, tracing the evolution of these frequency components in the current signal. The temporal evolution of the variation characteristics is aligned with the augmentation source's time stamp to ensure synchronous feature extraction; the sliding window is set according to the augmentation source's sampling rate. Feature correlation analysis uses the coupling relationship of the augmentation source as a reference template to identify similar coupling patterns in the current response. Physical classification assigns current features to corresponding categories based on the position, velocity, and acceleration information framework provided by the augmentation source. Position-related features include parameters such as displacement amplitude, trajectory shape, and center offset; velocity-related features include parameters such as rotational speed fluctuation, eddy frequency, and precession direction; and acceleration-related features include parameters such as vibration acceleration, impact response, and transient changes.

[0024] Step S120: Estimate rotor position based on current response change characteristics to extract rotor offset coordinate parameters, extract positioning accuracy drive source from the estimated deviation of rotor offset coordinate parameters, and generate impact-free start excitation current command based on current response change characteristics processed by positioning accuracy drive source.

[0025] Specifically, rotor position is estimated and rotor offset coordinate parameters are extracted based on the current response change characteristics. Based on the electromagnetic force balance principle, the radial offset distance is calculated according to the amplitude change parameters in the current change characteristics. The radial X-direction offset is calculated using the current difference between the positive and negative X coils: Δx = (iX+ - iX-) × Kpos, where iX+ is the X-positive coil current, iX- is the X-negative coil current, and Kpos is the position conversion coefficient (0.5 mm / A). The radial Y-direction offset is determined using the current difference between the positive and negative Y coils: Δy = (iY+ - iY-) × Kpos, where the positive and negative Y-directions represent the direction of deviation from the center. The axial Z-direction offset is calculated using the current ratio of the upper and lower thrust coils: Δz = (iup - idown) / (iup + idown) × gap, where iup is the upper thrust coil current, idown is the lower thrust coil current, and gap is the axial air gap length (2 mm). Position estimation considers the phase shift characteristics of the current response. Phase lead indicates that the rotor position lags behind the control command, requiring phase compensation correction. Frequency modulation characteristics are used to identify the rotor's dynamic offset pattern; modulation below 100Hz corresponds to slow drift, while 100-500Hz corresponds to rapid vibration. The time resolution of the offset coordinates reaches 50 microseconds, and discrete estimated points are connected into a continuous position trajectory through interpolation. The rotor offset coordinate parameters include complete information such as three-dimensional position coordinates (Δx, Δy, Δz), offset velocity vector, and offset acceleration. The estimated rotor offset coordinate parameters are compared with the desired position, and the estimation deviation ε = (ΔX - X0)² + (ΔY - Y0)² + (ΔZ - Z0)² is calculated, where ΔX, ΔY, and ΔZ are the previously calculated offsets, and X0, Y0, and Z0 are the desired positions (usually 0).

[0026] In some embodiments, extracting the positioning accuracy drive source from the estimated deviation of the rotor offset coordinate parameters includes: performing a sliding window analysis on the estimated deviation of the rotor offset coordinate parameters to obtain local deviation peaks; arranging the local deviation peaks in time sequence and performing gradient analysis to determine the deviation change trend; identifying deviation turning point features based on the slope abrupt change points of the deviation change trend; and generating the positioning accuracy drive source according to the time distribution density of the deviation turning point features.

[0027] A sliding window analysis was used to obtain local deviation peaks from the estimated deviations of the rotor offset coordinate parameters. The calculated estimated deviation value ε was analyzed locally using the sliding window method to identify peak points in the deviation sequence. The length of the sliding window was set to 200 sampling points, corresponding to a time span of 10 milliseconds, with a step size of 20 sampling points to ensure an 85% overlap rate. Peaks in the deviation data within the window were identified using an extreme value detection method; when the deviation value at a point was greater than the values ​​of the five points before and after it, it was marked as a local peak. The threshold for determining peaks was dynamically adjusted based on the statistical characteristics of the deviations within the window, set at the window mean plus twice the standard deviation. Radial deviation peaks reflect the extreme positions of rotor whirl; the periodicity of peak occurrence corresponds to the whirl frequency, while non-periodic peaks indicate the presence of random disturbances. Axial deviation peaks correspond to the extreme points of axial vibration; consecutive peaks indicate persistent axial instability. Peak amplitudes were graded into three levels: Level 1 peaks (>100 micrometers), Level 2 peaks (50-100 micrometers), and Level 3 peaks (20-50 micrometers). Each peak record includes its occurrence time, amplitude, duration, and corresponding coordinate axis. The spatial distribution of local deviation peaks is represented by a three-dimensional scatter plot. The peak distribution in the XY plane reflects radial positioning characteristics, while the peak distribution in the Z-axis reflects axial control characteristics.

[0028] Gradient analysis is performed to determine the trend of local deviation peaks after they are arranged in chronological order. The identified local deviation peaks are arranged in chronological order, and the trend of the peak sequence is analyzed by gradient calculation. The gradient analysis uses the central difference method, gi=(Pi+1-Pi-1) / (2Δt), where Pi is the amplitude of the i-th peak, Pi+1 and Pi-1 are adjacent peaks, Δt is the peak time interval, and gi is the gradient value. A positive gradient indicates that the deviation peak is increasing, and the system positioning accuracy is deteriorating; a negative gradient indicates that the deviation peak is decreasing, and the positioning accuracy is improving. The absolute value of the gradient reflects the rate of change; when the absolute value of the gradient exceeds 10 micrometers / second, it indicates rapid change, requiring timely adjustment of control parameters. Multiple consecutive positive gradients (more than 5) indicate that the system may be entering an unstable state, requiring preventative measures. Statistical analysis of the gradient sequence extracts trend characteristics, including the gradient mean, gradient variance, and gradient autocorrelation function. The gradient mean reflects the overall trend; a positive value indicates an overall increase in deviation, and a negative value indicates an overall decrease in deviation. The gradient variance reflects the severity of change; a larger variance indicates a more unstable system. The autocorrelation function identifies the periodicity of the gradient sequence, and the time delay corresponding to the correlation peak reflects the characteristic periodicity of the deviation change. The piecewise fitting of the change trend adopts the least squares method, dividing the gradient sequence into rising, stationary, and falling segments, and approximating each segment with a linear function.

[0029] Identifying deviation turning points based on slope abrupt change points in deviation trends. The deviation trend is determined by analyzing the gradient sequence, and key feature points indicating trend reversals are identified through slope abrupt change detection. The detection of slope abrupt change points employs a second-order difference method, identifying the abrupt change location by calculating the rate of change of the gradient; a rate of change exceeding 5 μm / s² is considered a turning point. Turning points divide the deviation evolution process into different stages, each corresponding to a different system state or control mode. Rising turning points indicate the beginning of a rapid increase in deviation, typically corresponding to the introduction of a disturbance or a degradation of control performance. Falling turning points indicate the beginning of a decrease in deviation, corresponding to the activation of control measures or the elimination of a disturbance. Successive turning points form an oscillation pattern; the oscillation period and amplitude reflect the dynamic characteristics of the system. The extraction of turning point features includes parameters such as turning angle, turning point intensity, and turning point duration. The turning angle is calculated by the difference in slope before and after the turning point; a larger angle indicates a more abrupt turning point. The turning point intensity is reflected by the curvature at the turning point; a larger curvature indicates a sharper turning point. The turning point duration is the time span from the start of the turning point to its completion, reflecting the system's response speed. Transition characteristics are classified according to their frequency of occurrence and degree of influence. High-frequency transitions reflect rapid adjustment processes, while low-frequency transitions reflect slow drift processes.

[0030] The positioning accuracy drive source is generated based on the temporal distribution density of deviation transition features. The temporal distribution density of deviation transition features is calculated using the kernel density estimation method. A Gaussian kernel function is placed at each transition point, and a continuous density distribution is obtained through superposition. High-density regions (greater than 10 transitions / second) indicate frequent transitions, corresponding to unstable periods or critical transition phases in the system. Low-density regions (less than 1 transition / second) indicate stable system operation and relatively stable positioning accuracy. Periodic analysis of density peaks identifies the system's inherent oscillation modes; oscillations with periods in the range of 0.1-1 seconds are usually related to the control loop. The construction of the drive source integrates density distribution, transition intensity, and trend information to form a multi-dimensional feature description. The primary drive component corresponds to the frequency component with the highest density, and the secondary drive component corresponds to the second-highest density region. The intensity of the drive source is quantized through density integral, and the intensity value is normalized to the 0-1 range: above 0.7 indicates strong drive, 0.3-0.7 indicates medium drive, and below 0.3 indicates weak drive. The generated positioning accuracy drive source contains complete information including temporal features, frequency features, intensity distribution, and phase relationships.

[0031] The shockless start-up excitation current command is generated based on the current response variation characteristics of the driving source with positioning accuracy. The key to shockless start-up lies in eliminating the identified deviation components in the driving source, especially low-frequency drift and transient impact components. The processing involves reverse compensation of the driving source's deviation amplitude spectrum. When a positioning deviation of 10 micrometers is detected at a certain frequency, a corresponding compensation component is superimposed on the excitation current. The amplitude of the compensation current is calculated as icomp = -ε × Kcomp × H(f), where ε is the deviation value, Kcomp is the compensation gain (0.02 A / μm), and H(f) is the frequency correlation transfer function. Phase compensation is achieved through phase inversion, ensuring that the compensation current and the deviation signal are 180 degrees out of phase, thus canceling the deviation. The current command during the start-up process is planned using an S-curve, with an initial slope of 0.1 A / s, a gradual increase in slope to 1 A / s in the middle section, and a final slope decreasing back to 0.1 A / s. The time parameters of the S-curve are determined based on the system response time identified by the driving source, and the total start-up time is set to 5-10 times the system time constant. The radial excitation current command consists of two parts: a base levitation current and a dynamic compensation current. The base current provides static levitation force, while the compensation current eliminates dynamic deviations. The axial excitation current command considers gravity preload; the initial current of the upper thrust coil is greater than that of the lower thrust coil, and the difference is equal to the current value corresponding to the rotor's gravity. The current change rate is monitored in real time during command generation and limited to within 2A / s to avoid mechanical shock caused by excessively rapid current changes.

[0032] Step S130: Based on the non-impact start excitation current command, monitor the operating status of multiple sensors to establish a fusion data network, activate the signal reconstruction configuration through the abnormal response in the fusion data network, and generate a degraded operation fault-tolerant mode based on the signal reconstruction configuration and the current response change characteristics.

[0033] Specifically, a fusion data network is established based on monitoring the operating status of multiple sensors using the excitation current command for shockless start-up. The multiple sensors include displacement sensors, acceleration sensors, temperature sensors, and current sensors, with multiple measuring points of each type deployed at different locations within the magnetic bearing system. The displacement sensor monitors the actual position change of the rotor after executing the excitation current command, comparing the measured displacement with the expected displacement; a deviation exceeding 5 micrometers indicates an anomaly. The acceleration sensor detects the vibration response during the start-up process; when the excitation current changes according to an S-shaped curve, the acceleration should remain below 0.1g; exceeding this limit indicates an impact. The temperature sensor monitors the coil temperature rise; the temperature rise rate during shockless start-up should be less than 2℃ / minute; excessively rapid temperature rise indicates abnormal losses. The current sensor tracks the following relationship between the actual current and the commanded current; the tracking error should be less than 2% of the commanded value; excessive error indicates a problem in the control loop. The fusion data network is established using a star topology, with the excitation current command as the central node and the data from each sensor as peripheral nodes. The connection weights between nodes are determined based on the sensor's sensitivity to command execution: displacement sensor weight 0.4, acceleration sensor weight 0.3, temperature sensor weight 0.2, and current sensor weight 0.1. Data fusion uses a weighted average method, with the fused value F = Σ(wi × si), where wi is the weight of the i-th sensor and si is the normalized value of the sensor measurement. The network's time synchronization accuracy reaches 100 microseconds, ensuring alignment of data from different sensors on the time axis. An abnormal response signal is generated when the fused value F deviates from the normal range or its rate of change exceeds a set threshold.

[0034] In some embodiments, the configuration reconstruction via anomaly response activation signal in the fused data network includes: detecting sensor drift patterns using anomaly response of the fused data network; performing multi-channel collaborative calibration to form a collaborative calibration trajectory sequence using the sensor drift patterns; performing spatial reconstruction mapping on the collaborative calibration trajectory sequence to obtain a set of sensor compensation coordinates; and forming a signal reconstruction configuration based on the set of sensor compensation coordinates.

[0035] Sensor drift patterns are detected using anomaly response analysis of a fused data network. Sensor drift detection is achieved by analyzing the long-term trend of the fused value F. Drift is identified when F monotonically changes beyond a set threshold over a continuous period. Drift patterns are categorized as linear and nonlinear. Linear drift is characterized by a constant rate of output deviation, while nonlinear drift exhibits a rate of change that varies over time. The slope of linear drift is determined using a least-squares fitting method, calculating the drift rate by dividing the covariance of each time point with the fused value by the time variance. Nonlinear drift is fitted using polynomials, typically quadratic or cubic polynomials, with coefficients determined by minimizing the fitting error. Typical drift scenarios include displacement sensors experiencing a slow increase in output due to probe surface contamination (characterized by positive linear drift); temperature sensors exhibiting nonlinear drift due to changes in thermistor characteristics caused by aging; and accelerometers experiencing periodic drift due to low-frequency vibrations caused by a loose base. The drift direction is identified by sign: positive drift indicates an increase in output value, while negative drift indicates a decrease. Multi-sensor drift correlation analysis identifies common-mode drift and differential-mode drift. Common-mode drift is usually caused by external factors such as changes in ambient temperature, and all sensors exhibit the same trend. Differential-mode drift is mostly caused by the aging of individual sensors, and different sensors have different drift directions and rates. Characteristic parameters of drift modes include drift onset time, drift rate, drift acceleration, and predicted failure time.

[0036] Multi-channel collaborative calibration is performed using sensor drift mode to form a collaborative calibration trajectory sequence. Collaborative calibration utilizes the linear or nonlinear characteristics of the drift mode to establish calibration relationships between adjacent channels, with the channel with the smaller drift serving as the reference. For linear drift, slope compensation is used: the calibration value Ci = Mi - k × (t - t0), where Mi is the measured value, k is the drift slope, t is the current time, and t0 is the drift start time. For nonlinear drift, polynomial compensation is used, calculating the compensation amount at each time step based on the fitted polynomial function. Multi-channel collaboration is achieved through cross-comparison; the calibration value of each channel is compared to the average value of other channels, and the deviation is used to correct the calibration parameters. The collaborative calibration trajectory is generated by recording the calibration values ​​of each channel in chronological order, forming a two-dimensional trajectory array, with rows corresponding to time points and columns corresponding to different channels. The sampling interval for the trajectory sequence is set to 1 second, continuously recording 100 points to cover the 100-second calibration process. The trajectory is smoothed using a moving average method with a window length of 5 sampling points to eliminate random fluctuations during the calibration process. The calibration effect is evaluated through residual analysis. A residual less than 10% of the original drift indicates that the calibration is effective.

[0037] Spatial reconstruction mapping is performed on the collaborative calibration trajectory sequence to obtain the sensor compensation coordinate set. Spatial reconstruction mapping projects the time-domain calibration trajectory onto a spatial coordinate system, with each trajectory point corresponding to a spatial compensation vector. The mapping relationship is achieved through a transformation matrix T. The coordinate set is defined as follows: Xc, Yc, and Zc are the compensation coordinates; C1-C4 are the calibration values ​​for the four channels; and T is a 4×3 transformation matrix. The transformation matrix is ​​determined based on the geometric arrangement of the sensors; orthogonally arranged sensors require orthogonal transformations, while non-orthogonal arrangements require oblique transformations. The physical meaning of the compensation coordinates corresponds to the correction amount of the rotor position: the X and Y directions compensate for radial position errors, and the Z direction compensates for axial position errors. The coordinate set is constructed by collecting the compensation coordinates at all times into a point cloud; the point cloud density reflects the time resolution of the calibration. The compensation range is determined through envelope analysis of the point cloud; the maximum compensation amount is typically limited to within 5% of the sensor's range. Spatial interpolation uses a trilinear interpolation method to generate a continuous compensation field between discrete compensation points. The data structure of the compensation coordinate set includes attributes such as coordinate values, time labels, effective range, and application priority.

[0038] A signal reconstruction configuration is formed based on a set of compensated coordinates from the sensors. The Xc, Yc, and Zc values ​​from the compensated coordinate set are used as correction parameters and applied to the original output of the corresponding sensors. The reconstruction rule is Srecon = Sraw + Sc, where Srecon is the reconstructed signal, Sraw is the original signal, and Sc is the compensation value extracted from the coordinate set. The compensation value is selected based on the current time and sensor position, finding the nearest compensation point from the set or calculating it through interpolation. Configuration schemes include a primary configuration and a backup configuration. The primary configuration uses the complete compensated coordinate set, while the backup configuration uses simplified linear compensation. Configuration parameter settings include a compensation enable flag, a compensation strength coefficient, and a compensation range limit. The strength coefficient is typically set between 0.8 and 1.0. Signal quality is evaluated by the improvement in signal-to-noise ratio (SNR) before and after compensation; an SNR improvement exceeding 3dB indicates effective reconstruction. The configuration takes effect under three conditions: sensor drift exceeding a threshold, drift duration exceeding a set value, and other sensor-confirmed anomalies. The reconstruction configuration is stored in a structured format, including management information such as configuration version, generation time, validity period, and application records. The configuration switching adopts a double buffering method. The new configuration is switched instantly after the background preparation is completed, so as to avoid the reconstruction process affecting real-time control.

[0039] Degraded operation fault-tolerant modes are generated based on signal reconstruction configuration combined with current response change characteristics. The degree of degradation is determined by the type of signal reconstruction configuration: parameter adjustment configuration corresponds to Level 1 degradation (10% performance reduction), sensor switching configuration corresponds to Level 2 degradation (30% performance reduction), and system reconfiguration configuration corresponds to Level 3 degradation (50% performance reduction). The fault-tolerant modes use key parameters from the current response change characteristics as supplementary information sources. When the displacement sensor fails, the position is deduced from the current change using the current-displacement conversion relationship. Level 1 degradation mode maintains the original control bandwidth, adjusting only the control parameters, reducing the proportional gain by 20% and increasing the integral time by 50% to ensure system stability. Level 2 degradation mode reduces the control bandwidth to 70% of its original value, decreasing dynamic response speed but improving robustness, and limiting the maximum speed to 80% of the rated value. Level 3 degradation mode adopts a conservative control strategy, reducing the control bandwidth to 50% of its original value, limiting the speed to 60% of the rated value, and increasing the air gap setpoint by 20% to provide a greater safety margin. The switching conditions for fault-tolerant modes include the number of sensor failures, the duration of the failures, and system performance indicators. Meeting any one of these conditions triggers the corresponding degradation mode. Mode transitions employ a smooth transition method, gradually adjusting parameters over 5 seconds to avoid transient disturbances caused by mode switching. Each degradation mode retains the ability to degrade to a lower level, ensuring the system maintains basic functionality even under multiple failure conditions.

[0040] Step S140: Receive the shutdown command signal through the degraded operation fault-tolerant mode, convert the shutdown command signal into a response preparation gain, and perform a shutdown state assessment based on the response preparation gain and the signal reconstruction configuration to generate emergency shutdown trigger conditions.

[0041] Specifically, shutdown command signals are received through a degraded operation fault-tolerant mode. The shutdown command reception strategy is determined based on the degrade level of the fault-tolerant mode: Level 1 maintains the normal command reception channel, Level 2 activates redundant reception channels, and Level 3 employs multi-channel voting reception. Shutdown command signals include three types: normal shutdown, rapid shutdown, and emergency shutdown, each with different priorities and execution sequences. Normal shutdown commands are received through the main control channel in fault-tolerant mode, with a signal format of a continuous high level for more than 100 milliseconds, indicating a planned shutdown operation. Rapid shutdown commands are received simultaneously through the main control and backup channels, with a double-pulse sequence and a pulse interval of 50 milliseconds, indicating a shutdown must be completed within 30 seconds. Emergency shutdown commands are broadcast through all available channels, with a continuous square wave at a frequency of 10Hz, indicating immediate shutdown execution. The fault-tolerant mode performs validity checks on received commands: in Level 1 degrade, the signal strength is required to reach 90% of the threshold; in Level 2 degrade, this is reduced to 70%; and in Level 3 degrade, it is reduced to 50%. The timestamps for received instructions are accurate to the millisecond level, recording the instruction arrival time, acknowledgment time, and start execution time. Instructions received from multiple channels are aligned using a 10-millisecond time window to ensure that the same instruction from different channels is correctly identified.

[0042] In some embodiments, converting the shutdown command signal into a response preparation gain includes: decomposing the shutdown command signal into a dynamic delay sequence by time window; performing buffer state analysis on the dynamic delay sequence to generate a buffer time matrix; performing phase matching between the dynamic delay sequence and the buffer time matrix to form a timing buffer map; and extracting preparation enhancement nodes from the timing buffer map to determine the response preparation gain.

[0043] The shutdown command signal is decomposed into a dynamic delay sequence by time windows. The received shutdown command signal is segmented into time windows, breaking down the continuous signal into sequence segments with different delay characteristics. The time window setting is determined based on the characteristics of the command signal: a 100-millisecond window for normal shutdown commands, a 50-millisecond window for rapid shutdown, and a 20-millisecond window for emergency shutdown. Window decomposition uses a sliding method, with adjacent windows overlapping by 50% to ensure the continuity of signal characteristics is not disrupted. The start time, peak time, and end time of each signal segment within a window are extracted to form a time characteristic triplet for that segment. Dynamic delay calculation compares the actual arrival time of each signal segment with the ideal arrival time; the delay value equals the actual time minus the ideal time. The delay sequence is arranged according to the window order, forming a time sequence reflecting the delay of command transmission and processing. The delay values ​​in the sequence exhibit dynamic changes: the initial segment has a larger delay (10-20 milliseconds), the middle segment has a stable delay (5-10 milliseconds), and the final segment may have an increased delay. Abnormal delays are identified by comparing with the average delay; delays exceeding twice the average are marked as abnormal points. The length of the dynamic delay sequence is equal to the number of segments in the window decomposition, typically 10-50 segments, depending on the duration of the command signal. The sequence storage includes attribute information such as delay value, window index, signal strength, and anomaly flags.

[0044] A buffer state analysis is performed on the dynamic delay sequence to generate a buffer time matrix. The buffer state analysis checks the maximum buffer time that each delay segment in the dynamic delay sequence can accommodate. The buffer time is equal to the system response deadline minus the current accumulated delay. The buffer capacity Bi of the i-th segment is Bi = Tdeadline - Σ(Dj), where Tdeadline is the downtime limit, Dj is the delay of the j-th segment, and the summation ranges from 1 to i. The construction of the buffer time matrix organizes the buffer capacity of each segment into a two-dimensional structure, with rows corresponding to delay sequence segments and columns corresponding to different buffering strategies. The first column of the matrix is ​​the basic buffer time, which is directly equal to the calculated buffer capacity value of each segment. The second column is the safety buffer time, reserving a 20% safety margin on top of the basic buffer. The third column is the limit buffer time, considering the maximum processing capacity of the system, which is usually 1.5 times the basic buffer. The buffer state is classified according to the adequacy of the buffer time: sufficient state (>100 milliseconds), normal state (50-100 milliseconds), strained state (20-50 milliseconds), and critical state (<20 milliseconds). Negative values ​​in the matrix indicate that the segment has exceeded the buffer capacity and needs to be processed immediately or certain steps should be skipped. The buffer time matrix has a dimension of N×3, where N is the length of the delay sequence and 3 corresponds to the three buffering strategies. The matrix is ​​updated in real time, with the buffer state of subsequent segments updated immediately after each segment is processed.

[0045] For example, the step of performing phase matching between the dynamic delay sequence and the buffer time matrix to form a timing buffer map includes: converting the dynamic delay sequence into a virtual oscillation signal; allowing the virtual oscillation signal to propagate in the buffer time matrix to generate an interference pattern; extracting enhancement nodes and reduction nodes in the interference pattern; and encoding the spatial distribution of the enhancement nodes and the reduction nodes into a timing buffer map.

[0046] The dynamic delay sequence is converted into a virtual oscillating signal. The calculated delay values ​​(the difference between actual and ideal time) in the dynamic delay sequence are mapped to a periodic oscillating signal to capture the dynamic characteristics of the delay. The conversion process maps the delay value Di of the i-th segment of the sequence to an amplitude Ai = A0 × (Di / Davg), where A0 is the reference amplitude (set to 1), and Davg is the average of all delay values. This linear mapping preserves the relative magnitude of the delays. The oscillation frequency is determined based on the rate of change of the delay. Segments with a large rate of change ΔDi / Δt correspond to high-frequency components (indicating rapid changes), where ΔDi is the difference between adjacent delay values ​​and Δt is the time interval. Segments with a small rate of change correspond to low-frequency components (indicating slow changes). The phase of the virtual oscillating signal is calculated based on the delay accumulation effect: φi = 2π × (ΣDj / ΣDmax), where φi is the phase of the i-th segment, ΣDj is the cumulative delay sum from segment 1 to segment i, and ΣDmax is the maximum possible cumulative delay sum, reflecting the degree of delay accumulation. The signal envelope is obtained by connecting local extrema. The upper envelope corresponds to the evolution trend of the delay peak, and the lower envelope corresponds to the evolution trend of the delay trough. The modulation characteristics of the oscillating signal reflect the abrupt changes in delay; a modulation abrupt change marker is generated when the change between adjacent delay values ​​exceeds 50%. The complete representation of the signal includes four components: amplitude sequence, rate of change sequence, cumulative phase sequence, and envelope function. The temporal resolution of the virtual signal is consistent with the original sequence, with each delay value corresponding to one oscillation sampling point. The generated virtual oscillating signal, as another representation of the delay sequence, retains all temporal information.

[0047] A virtual oscillating signal is propagated within a buffer time matrix to generate an interference pattern. The virtual oscillating signal is used as input, and its propagation is simulated in the computational space defined by the buffer time matrix. During propagation, each element Bij of the matrix is ​​considered a propagation coefficient. The output of the oscillating signal at that position is Oij = Ai × Bij, where Oij is the output signal strength in the i-th row and j-th column, Ai is the oscillation amplitude in the i-th segment, and Bij is the buffer time value in the i-th row and j-th column of the buffer time matrix. The propagation coefficient Bij reflects the buffering capacity at that moment; the larger the buffer time, the larger the propagation coefficient, and the more complete the signal transmission. The oscillating signal propagates downwards row by row from the first row of the matrix, generating a new output signal at each row. When the signal encounters a point where the buffer time changes, the output signal bifurcates: one part continues to propagate (transmission component), and the other part returns (reflection component). The transmission component Tij = Oij × (Bij / Bij-1)^0.5, and the reflection component Rij = Oij × (1-Bij / Bij-1)^0.5, where Tij is the transmission intensity, Rij is the reflection intensity, and Bij-1 is the buffer time of the previous row and column. When Bij > Bij-1, transmission is dominant; otherwise, reflection is dominant. When the forward propagating signal and the reflected signal are superimposed at the same position, an interference effect occurs. The superposition result Iij = Tij + Rij × cos(φi-φj), where Iij is the total interference intensity, φi and φj are the phases of the two signals, and cos is the cosine function. The sign of the interference intensity depends on the phase relationship; when in phase (positive cosine value), it enhances the signal, and when out of phase (negative cosine value), it weakens the signal. The interference pattern forms a strong-weak distribution in the matrix space, with strong interference regions corresponding to periods of coordinated system response and weak interference regions corresponding to periods of conflicting response.

[0048] Enhanced and ablation nodes are extracted from the interferogram. Enhanced nodes are identified using a threshold method: when the interferometric intensity Iij is greater than the average intensity plus one standard deviation, it is marked as an enhanced node. Ablation nodes are identified when the interferometric intensity is less than the average intensity minus one standard deviation; these locations indicate weaker system response. Node intensity levels are classified based on their deviation from the mean: more than two standard deviations indicate strong nodes, one to two standard deviations indicate moderate nodes, and less than one standard deviation indicates weak nodes. Spatial distribution characteristics of nodes include node density (number of nodes / total grid), average spacing, and clustering. Connectivity analysis of enhanced nodes identifies node clusters; continuous enhanced nodes form enhanced channels, indicating a sustained period of high response. Distribution patterns of ablation nodes identify weak points in the system; clustered ablation nodes require focused attention and compensation. Temporal stability of nodes is assessed by comparing their states at adjacent times; nodes whose states remain unchanged are stable, while those that change frequently are oscillating nodes. Identification of critical nodes is based on their influence range, quantified by the change in the interferogram after removing the node. The statistics of node characteristics include indicators such as the number, proportion, and distribution variance of various types of nodes.

[0049] The spatial distribution of augmenting and ablation nodes is encoded into a temporal buffer graph. Augmenting nodes are encoded according to their strength level: strong augmentation is encoded as 3, moderate augmentation as 2, and weak augmentation as 1, reflecting the node's contribution to the system response. Ablation nodes are encoded with negative values: strong ablation as -3, moderate ablation as -2, and weak ablation as -1, indicating a weakening effect on the system response. Neutral regions (neither augmenting nor ablation) are encoded as 0, indicating that the buffer state at that location is balanced. Spatial location encoding uses two-dimensional coordinates, with each encoded point represented as a triple (i,j,Cij), where i is the row number, j is the column number, and Cij is the encoded value at that location (an integer between -3 and +3). The temporal buffer graph data structure uses a sparse representation, recording only non-zero encoded points to reduce storage requirements and highlight key information. The temporal dimension of the graph is represented by a third coordinate axis, forming a three-dimensional spatiotemporal encoding (i,j,t,Cij), where t is the time index. The coding density ρ = Nnonzero / Ntotal reflects the activity level of the system, where ρ is the coding density, Nnonzero is the number of non-zero coding points, and Ntotal is the total number of grid points. High density indicates frequent changes in the buffer state. The spatial autocorrelation function of the coding calculates the correlation between adjacent coding points; strong correlation indicates spatial continuity of the buffer state. The integrity of the temporal buffer map is ensured by an interpolation algorithm, filling in missing coding points using nearest neighbor interpolation or linear interpolation. The generated temporal buffer map includes three parts: the coding matrix, statistical features, and metadata.

[0050] The response readiness gain is determined by extracting boosting nodes from the time-series buffer graph. Boosting nodes are identified by searching for peak buffer capacity points in the time-series buffer graph, corresponding to the moments when the system's response capability is strongest. Node selection criteria include a buffer time greater than 1.5 times the average, a duration exceeding 20 milliseconds, and stable state before and after the buffer. The strength of a boosting node is quantified by its prominence in the graph; the strength value equals the ratio of that node's buffer capacity to the surrounding average. The response readiness gain is calculated by integrating the contributions of all boosting nodes: G = Σ(Si × Wi) / Σ(Wi), where Si is the strength of the i-th node and Wi is the node weight. Node weights are determined based on their importance during the shutdown process: 0.3 for early nodes, 0.5 for mid-term nodes, and 0.2 for late-term nodes. Gain value adjustments consider the uniformity of node distribution: a reward coefficient of 1.1 is applied when node distribution is uniform, and a penalty coefficient of 0.9 is applied when distribution is concentrated. The final response readiness gain is standardized to the range of 0-2, where 0 represents no readiness, 1 represents normal readiness, and 2 represents adequate readiness. The gain value includes a credibility score, which is based on the number and quality of the enhanced nodes. The more nodes and the higher the strength, the higher the credibility.

[0051] Emergency shutdown trigger conditions are generated based on the response preparation gain combined with the signal reconstruction configuration for shutdown state assessment. The response preparation gain is used as a weighting factor to weight the various parameters in the signal reconstruction configuration. When the gain value is greater than 1.8 and the reconstruction configuration is in a system reconfiguration state, the assessment result tends to be an immediate shutdown. The assessment indicators include four aspects: current rotor speed, vibration amplitude, bearing temperature, and control margin. The weight of each indicator in the assessment is determined according to the gain value. The speed assessment considers the time required to reduce the current speed to zero. The higher the gain, the shorter the allowable deceleration time. A gain of 2.0 requires a shutdown within 10 seconds. The vibration assessment checks whether the current vibration level allows for a normal shutdown process. If the vibration exceeds 80% of the allowable value, the shutdown strategy needs to be adjusted. The temperature assessment determines whether there is a risk of bearing overheating. If the temperature exceeds the alarm value, a rapid cooling procedure is triggered. The control margin assessment judges the compensation capability of the reconstruction configuration. If the coverage of the compensation coordinate set is less than 50% of the requirement, the margin is considered insufficient. The emergency stop trigger conditions are generated based on a comprehensive evaluation of various factors, with three trigger thresholds set: any two indicators exceeding limits, a sudden increase in gain exceeding 0.5, and configuration rebuild failure. Trigger conditions include parameters such as trigger type, trigger threshold, delay time, and execution priority. The logical combination of conditions uses an "OR" relationship; the emergency stop procedure is initiated when any condition is met.

[0052] Step S150: Detect the power supply status of the bus capacitor based on the emergency stop trigger condition, release the gradient control potential from the energy change of the bus capacitor power supply status, determine the active damping control timing based on the gradient control potential combined with the degraded operation fault tolerance mode, and generate a damping force control command through the active damping control timing to realize intelligent start-stop control of the magnetic levitation refrigeration compressor.

[0053] Specifically, the bus capacitor power supply status is detected based on the emergency shutdown trigger condition. The detection of the bus capacitor power supply status includes four key parameters: capacitor voltage, discharge current, energy storage capacity, and internal resistance change. The capacitor voltage is measured using a high-precision voltage sensor; the normal operating voltage is 540V, and the voltage drop curve is recorded when the emergency shutdown trigger condition is activated. The discharge current monitors the instantaneous current supplied by the capacitor to the magnetic bearing system; the current is in the range of 10-20A during normal shutdown and may reach 30-50A during emergency shutdown. The energy storage capacity is obtained through integral calculation: E = 0.5 × C × V², where C is the capacitance value (typically 10000μF), V is the current voltage, and E is the stored energy. The internal resistance change reflects the health status of the capacitor, calculated as the ratio of voltage drop to current; an increase in internal resistance indicates capacitor aging. The detection frequency is adjusted according to the urgency of the trigger condition; the sampling frequency is 1kHz during normal triggering and increased to 10kHz during emergency triggering. The power supply status is graded into four levels: sufficient (voltage > 500V), normal (450-500V), warning (400-450V), and critical (< 400V). Abnormal status is identified by voltage drop (drop rate > 50V / s), current spike (peak value > 80A), and oscillating discharge (voltage fluctuation > 10V).

[0054] In some embodiments, releasing gradient control potential from energy changes in the power supply state of the bus capacitor includes: monitoring voltage decay gradient patterns using the power supply state of the bus capacitor; performing capacitor discharge optimization control through the voltage decay gradient patterns to form an optimized discharge region; implementing graded current regulation on the optimized discharge region to obtain a power supply current directional distribution; and forming gradient control potential based on the power supply current directional distribution.

[0055] Voltage decay gradient patterns are monitored using the bus capacitor power supply status. By analyzing the voltage change characteristics during bus capacitor power supply, the gradient change pattern of voltage decay over time is identified. Voltage decay monitoring begins at the emergency shutdown trigger moment, recording voltage values ​​every millisecond to form a high-density voltage time series. The decay gradient is calculated using a five-point difference method to improve the accuracy and noise immunity of gradient calculation. Gradient patterns are classified into three basic types: linear decay, exponential decay, and step decay. Linear decay is characterized by a constant voltage drop rate, typically occurring under constant load conditions, with a decay rate of 20-30V / s. Exponential decay follows the law V(t) = V0 × exp(-t / τ), where V0 is the initial voltage and τ is the time constant, reflecting the natural discharge characteristics of the RC circuit. Step decay corresponds to graded load switching, with each load level corresponding to a different decay rate, and a plateau period between levels. Hybrid modes identify combinations of multiple modes during the actual decay process; initially, it may be exponential decay, then it transitions to linear decay in the middle stage, and finally exhibits step decay characteristics. The detection of abnormal decay modes includes oscillating decay (voltage fluctuations), abrupt decay (voltage drops suddenly), and anomalous rise (voltage rises temporarily). Mode characteristic parameters extracted include average decay rate, decay acceleration, mode transition point, and decay stability.

[0056] Optimized discharge regions are formed by implementing capacitor discharge optimization control through voltage decay gradient modes. Differentiated discharge control is implemented based on the identified voltage decay gradient modes: linear decay mode employs a constant current discharge strategy, maintaining a constant current of 20A to ensure stable power supply; exponential decay mode uses variable current discharge, initially responding with a high current of 40A, then gradually reducing to 10A to maintain stability; step decay mode adjusts discharge parameters at each voltage step, adding current compensation at step intervals. Optimized regions are divided based on the discharge characteristics of different gradient modes: the high-voltage region (>480V) corresponds to constant current control in linear mode, the medium-voltage region (420-480V) is adapted to variable current regulation in exponential mode, and the low-voltage region (360-420V) uses segmented control in step mode. The region boundaries are dynamically adjusted according to the gradient mode; the boundaries shift forward in exponential decay mode and are evenly distributed in linear decay mode. Optimization objectives include minimizing downtime, ensuring smooth downtime, and reserving safety energy reserves. The discharge path selection prioritizes powering critical loads, with magnetic bearing control having the highest priority, followed by auxiliary systems. The area switching uses predictive control, which predicts the voltage change trend 5 seconds in advance and smoothly adjusts the discharge parameters.

[0057] A graded current regulation system is implemented in the optimized discharge region to obtain the current distribution. This graded current regulation further subdivides each optimized discharge region into 3-5 current levels, each corresponding to a specific current range and control objective. The first level (basic level) provides the minimum current to maintain levitation, ensuring the rotor does not contact the spare bearing; the current range is 5-10A. The second level (control level) provides the current required for position adjustment, achieving precise rotor position control; the current range is 10-20A. The third level (damping level) provides vibration suppression current, actively attenuating rotor vibration; the current range is 15-25A. The fourth level (braking level) provides deceleration braking current, rapidly reducing rotor speed; the current range is 20-35A. The current distribution is obtained through a current sensor array, monitoring the current flow from the bus capacitor to each load branch. The radial magnetic bearing receives 40-50% of the total current to maintain radial stability. The axial magnetic bearing receives 20-30% to balance axial force and gravity. The control system and sensors receive 10-15% to ensure normal control functions. The auxiliary systems receive the remaining 10-20%, including cooling and monitoring functions.

[0058] The gradient control potential is formed based on the current supply direction distribution. The calculation of the gradient control potential is based on the current supply direction distribution characteristics of each branch: radial branches carry 40-50% of the total current, corresponding to an adjustable control potential of 20A; axial branches carry 20-30% of the total current, corresponding to a control potential of 15A; cooling branches carry 15-20% of the total current, corresponding to a control potential of 10A; and auxiliary branches carry 5-10% of the total current, corresponding to a control potential of 5A. The control potential Pi of each branch is determined according to the flow direction distribution ratio and current regulation capability. The total control potential P = Σ(Pi × ηi), where ηi is the control efficiency coefficient of each branch, with radial branch efficiency of 0.9, axial branch efficiency of 0.8, and cooling branch efficiency of 0.7. The time-varying characteristics of the potential decrease as the capacitor voltage decreases, and the potential decay rate is proportional to the voltage decay rate. The potential allocation strategy prioritizes key functions: 50% of the potential is allocated to suspension control, 30% to vibration suppression, and 20% to speed control. Potential reserves are managed by allocating 20% ​​of the total potential as an emergency reserve to cope with sudden disturbances or control deviations. The gradient direction is determined based on the difference between the current state and the target state, prioritizing the release of potential. Potential assessment indicators include four dimensions: controllable range, response speed, duration, and stability margin.

[0059] The timing of active damping control is determined based on gradient control potential combined with the fault-tolerant mode of degraded operation. The determination of the active damping control timing comprehensively considers two core factors: the energy gradient level and the fault-tolerant mode level. When the gradient control potential P > 30A and the system is in Level 1 fault-tolerant mode, standard damping control is initiated; when P is in the range of 15-30A and in Level 2 fault-tolerant mode, enhanced damping control is initiated; when P < 15A and in Level 3 fault-tolerant mode, emergency damping control is initiated. Timing determination also considers potential reserves, reserving 20% ​​of the control potential for emergency adjustment, with the actual usable potential being 80% of the total potential. The fault-tolerant mode level affects the timing selection strategy: in Level 1 degraded mode, damping is applied 10 milliseconds before the vibration peak; in Level 2 degraded mode, it is applied 20 milliseconds earlier; and in Level 3 degraded mode, it is applied 30 milliseconds earlier. The timing window is set considering the system response delay, with a typical window width of 50 milliseconds, within which the optimal intervention point is found. The criteria for determining the optimal intervention point include two requirements: the energy gradient is within the controllable range and the control margin is sufficient. Multi-objective optimization determines the optimal timing, with the objective function including vibration suppression effect, energy consumption, and downtime. The timing is dynamically adjusted based on real-time feedback; if vibration is not effectively suppressed, the timing is advanced by 5 milliseconds and the attempt is repeated. The duration of damping control is determined based on vibration decay; damping is gradually reduced when vibration drops to 120% of the target value, and removed when it drops to 105%.

[0060] In some embodiments, generating a damping force control command through the active damping control timing includes: monitoring the torque oscillation mode of the active damping control timing to generate a torque oscillation trajectory; injecting an anti-phase force at the peak position of the torque oscillation trajectory to form a torque standing wave; extracting deterministic torque components using the stable nodes of the torque standing wave; and superimposing the deterministic torque components in phase to generate a damping force control command.

[0061] The torque oscillation pattern generated by monitoring the timing of active damping control generates a torque oscillation trajectory. At the determined active damping control timing point, the torque oscillation characteristics of the rotor system are monitored. Torque oscillation is monitored by calculating the differential current of the radial magnetic bearing: torque M = (iX+ - iX-) × r, where M is the resultant control torque, iX+ is the current in the positive X coil, iX- is the current in the negative X coil, and r is the lever arm, i.e., the rotor radius, typically 0.05-0.08 m. The current difference (iX+ - iX-) generates a radial electromagnetic force, which, when multiplied by the lever arm r, forms the control torque. The oscillation pattern is identified and analyzed to understand the torque variation over time, including oscillation frequency, amplitude, and phase characteristics. The dominant oscillation frequency is usually equal to the rotor rotational frequency, with an amplitude ranging from 0.5 to 2.0 N·m, and the phase is related to the rotor's angular position. Secondary oscillation components include higher harmonics such as the second and third harmonics, reflecting the effects of rotor imbalance and shape deviations. The generation of the torque oscillation trajectory involves connecting torque values ​​sequentially over time to form a two-dimensional trajectory curve, with time on the horizontal axis and torque value on the vertical axis. The envelope of the trajectory reflects the changing trend of oscillation energy; the upper envelope connects peak points, and the lower envelope connects trough points. Oscillation modes are classified into stable oscillations (constant amplitude), growing oscillations (increasing amplitude), decaying oscillations (decreasing amplitude), and chaotic oscillations (irregular). Trajectory feature points are marked with peak points, trough points, zero-crossing points, and inflection points; these feature points determine the location where the damping force is applied. Statistical analysis of the oscillation period calculates the average period, standard deviation of the period, and the trend of period variation.

[0062] A standing wave is formed by injecting an antiphase torque at the crest of the torque oscillation trajectory. The crest locations in the torque oscillation trajectory are identified, and opposite-phase torques are precisely injected at these locations to create a standing wave effect. The precise location of the crest is determined by the first derivative of the trajectory; the point where the derivative changes from positive to negative is the crest. The antiphase torque is calculated as the negative value of the crest torque: Manti = -Mpeak × α, where α is the injection coefficient (0.6-0.8) to avoid overcompensation. The injection timing is controlled to account for system response delay, starting injection 5-10 milliseconds before the crest to ensure the antiphase torque reaches its maximum value at the crest. The formation mechanism of the standing wave is the superposition of the original oscillation and the injected antiphase torque. At specific locations, they cancel each other out, forming nodes; at other locations, they may amplify, forming antinodes. The characteristics of the standing wave mode include fixed node locations, stable antinode amplitudes, and a regular energy distribution. The node location corresponds to the point of minimum vibration and is the most stable operating point of the system, preferably used as the target shutdown location. The positions of antinodes need to be avoided to prevent the vibration from being amplified when stopping the machine at these locations.

[0063] For example, the step of extracting deterministic torque components using the stable nodes of the torque standing wave includes: performing oscillation energy analysis on the torque standing wave to obtain oscillation stability data; identifying torque optimization nodes and control optimization nodes in the oscillation stability data; performing parameter fusion based on the torque optimization nodes and the control optimization nodes to generate a torque coordination curve; and performing feature extraction analysis on the torque coordination curve to form deterministic torque components.

[0064] Oscillation energy analysis is performed on torque standing waves to obtain oscillation stability data. The energy distribution characteristics in the torque standing wave are analyzed. The standing wave is formed by the superposition of the original oscillating torque and the anti-phase damping torque: standing wave torque Mw = Morig + Manti, where Morig is the original oscillating torque and Manti is the injected anti-phase torque. The standing wave energy is calculated as Ew = ∫Mw²dt, and the total energy of the standing wave is obtained through integration. The node positions in the standing wave correspond to the minimum energy value, the antinode positions correspond to the maximum energy value, and the node spacing reflects the stable period of the standing wave. Energy distribution analysis divides the standing wave space into multiple segments, and the energy density of each segment is calculated. Segments with high density correspond to energy concentration areas. Stability indicators include energy volatility (standard deviation / mean), energy decay rate (energy reduction per unit time), and energy distribution uniformity. An energy volatility of less than 10% indicates stable oscillation, 10-20% is critically stable, and greater than 20% is unstable. The instantaneous energy is calculated using Hilbert transform to obtain the envelope, and the square of the envelope is the instantaneous energy density. Energy flow analysis traces the energy transfer path between nodes and antinodes of a standing wave, with the energy flow of a stable standing wave forming a closed loop. The frequency domain energy spectrum is obtained through power spectral density analysis, showing the distribution ratio of energy at each frequency component. A dominant frequency energy share exceeding 70% indicates a single, stable oscillation mode, while dispersed energy across multiple frequencies indicates complex coupling. The stability data is organized into three parts: time-domain stability indices, frequency-domain energy distribution, and spatial energy density maps.

[0065] Torque-optimized nodes and control-optimized nodes were identified in the oscillation stability data. Node identification comprehensively utilized three components of the stability data: identifying stable periods with torque amplitude changes of less than 5% from the time-domain stability index, these periods corresponded to potential torque-optimized nodes; locating frequency points where the dominant energy is concentrated from the frequency-domain energy distribution, with frequencies within ±2Hz of the dominant frequency corresponding to control-optimized nodes; and determining the locations of energy density minima from the spatial energy density map, these locations were marked as optimal control intervention points. Cross-validation of the three types of data ensured the accuracy of node identification. Spatial distribution analysis of the nodes showed that torque-optimized nodes were mainly located near standing wave nodes, with intervals of approximately half a wavelength. Control-optimized nodes partially overlapped with torque-optimized nodes, but were more concentrated at energy flow convergence points. The temporal characteristics of the nodes showed periodic occurrence, with 2-4 torque-optimized nodes and 1-2 control-optimized nodes appearing within each oscillation cycle. The node strength was graded into major nodes (top 20% of stability index), minor nodes (20-50%), and weak nodes (50-80%). The primary node is the priority control target, secondary nodes are backups, and weak nodes are used only when necessary. Node correlation analysis identifies the coupling relationships between nodes; the control of certain nodes can affect the stability of other nodes. The effective window for optimizing a node is defined as the range of time within which node characteristics remain excellent, typically with a window width of 20-50 milliseconds.

[0066] A torque coordination curve is generated by parameter fusion based on torque optimization nodes and control optimization nodes. Parameter fusion employs a weighted combination method, with torque optimization node parameters having a weight of 0.6 and control optimization node parameters having a weight of 0.4. The fusion rule assigns the maximum weight at overlapping nodes and distributes weights inversely proportional to distance at non-overlapping nodes. The torque coordination curve is generated by connecting the fused nodes using spline interpolation to ensure the curve's second-order continuity. The curve's shape characteristics include peak-valley distribution, slope variation, and curvature features; smooth segments correspond to stable control regions, while steep segments correspond to rapid adjustment regions. The coordination principle prioritizes system stability, pursuing optimal torque under stable conditions; in case of conflict, the stability weight is increased to 0.7. The curve's dynamic adjustment is corrected based on real-time feedback; if oscillations intensify, the weight of the control optimization node is increased; if the response is slow, the weight of the torque optimization node is increased. The curve is segmented into a starting segment, a transition segment, a stable segment, and a closing segment, each employing a different coordination strategy. The curve's boundary conditions ensure that the torque values ​​at the starting and ending points match the current system state, avoiding abrupt changes.

[0067] Feature extraction analysis is performed on the torque coordination curve to form deterministic torque components. Feature extraction employs multi-scale analysis to identify characteristic patterns of the curve at different time scales. Large-scale features (>100 milliseconds) reflect the overall trend, and the trend line is extracted as the basic torque component. Mid-scale features (10-100 milliseconds) correspond to the main oscillation modes, and the dominant frequency and its harmonic components are extracted. Small-scale features (<10 milliseconds) contain rapid transients, and high-frequency components beneficial to control are selectively extracted. Determinism is determined through repeatability testing; features that remain unchanged over multiple cycles have high determinism. The parameterized representation of the components includes four parameters: amplitude, frequency, phase, and attenuation coefficient, fully describing the component characteristics. The dominant components are identified by ranking them by energy contribution, selecting the top few components with a cumulative contribution of 80%. The physical interpretation of the components is related to specific control actions: low-frequency components correspond to stiffness compensation, mid-frequency components to damping provision, and high-frequency components to disturbance suppression. Independence checks of the components ensure that the correlation between each component is less than 0.3, avoiding redundancy and coupling. The final deterministic torque component set contains 5-8 independent components, covering a frequency range from DC to 5th harmonics.

[0068] The deterministic torque components are superimposed in phase to generate the damping force control command. Vector synthesis is used, and each component is represented as... The form is Mtotal = Σ(Mk×cos(ωkt+φk)), where Mk is the amplitude and φk is the phase angle. The superposition formula is Mtotal = Σ(Mk×cos(ωkt+φk)), where ωk is the angular frequency of the k-th component and t is time. DC components are directly superimposed to provide constant damping force, fundamental frequency components are superimposed synchronously according to rotor phase, and harmonic components are superimposed according to their respective frequencies and phases. The damping coefficient is embedded by multiplying the superposition result by the damping coefficient Cd, Fd = Cd × Mtotal, where Cd ranges from 0.1 to 0.3. The control command conversion converts the torque command into a current command, i = Fd / (Km×r), where Km is the torque coefficient and r is the lever arm length. Command limiting ensures that the current does not exceed the safe range, with the maximum current limited to 150% of the rated value. Time discretization samples continuous commands according to the control cycle at a sampling frequency of 10kHz to ensure the capture of all important dynamic characteristics. Smoothing of the command sequence uses low-pass filtering, with the cutoff frequency set to twice the highest control frequency to eliminate high-frequency noise. The final command includes four-way directional control current and two-way axial control current, each with independent amplitude and phase parameters, realizing intelligent start-stop control of the magnetic levitation refrigeration compressor.

[0069] To implement the intelligent start-stop control method for the magnetic levitation refrigeration compressor corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects. See also Figure 2 , Figure 2This diagram illustrates a structural block diagram of an intelligent start-stop control system 200 for a magnetic levitation refrigeration compressor according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The intelligent start-stop control system 200 for a magnetic levitation refrigeration compressor provided in this embodiment includes: The data acquisition module 201 is used to acquire the current response data of the magnetic bearing coil, perform fluctuation analysis on the current response data to obtain the position sensing enhancement source, and obtain the current response change characteristics based on the position sensing enhancement source. The position estimation module 202 is used to estimate the rotor position and extract rotor offset coordinate parameters from the current response change characteristics, extract the positioning accuracy drive source from the estimation deviation of the rotor offset coordinate parameters, and process the current response change characteristics based on the positioning accuracy drive source to generate a shockless start excitation current command. The fault monitoring module 203 is used to monitor the operating status of multiple sensors based on the non-impact start excitation current command, establish a fusion data network, reconstruct the configuration by activating the abnormal response signal in the fusion data network, and generate a degraded operation fault-tolerant mode based on the signal reconstruction configuration and the current response change characteristics. The shutdown control module 204 is used to receive a shutdown command signal through the degraded operation fault-tolerant mode, convert the shutdown command signal into a response preparation gain, and perform shutdown state assessment based on the response preparation gain and the signal reconstruction configuration to generate emergency shutdown trigger conditions. The damping execution module 205 is used to detect the power supply status of the bus capacitor according to the emergency stop triggering condition, release gradient control potential from the energy change of the power supply status of the bus capacitor, determine the active damping control timing based on the gradient control potential and the degraded operation fault tolerance mode, and generate a damping force control command through the active damping control timing to realize intelligent start-stop control of the magnetic levitation refrigeration compressor.

[0070] The intelligent start-stop control system 200 for the magnetic levitation refrigeration compressor described above can implement the intelligent start-stop control method for the magnetic levitation refrigeration compressor in the above method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining contents of this application embodiment can be referred to the contents of the above method embodiments, and will not be repeated in this embodiment.

[0071] The purpose of the above embodiments is to reproduce and derive the technical solution of the present invention by way of example, and to fully describe the technical solution, purpose and effect of the present invention. The purpose is to enable the public to have a more thorough and comprehensive understanding of the disclosure of the present invention, and not to limit the scope of protection of the present invention.

[0072] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for intelligent start-stop control of a magnetic levitation refrigeration compressor, characterized in that, include: The current response data of the magnetic bearing coil is collected, and the fluctuation analysis of the current response data is performed to obtain the position sensing enhancement source. Based on the position sensing enhancement source, the current response change characteristics are obtained. The rotor position is estimated and rotor offset coordinate parameters are extracted from the current response change characteristics. The positioning accuracy drive source is extracted from the estimation deviation of the rotor offset coordinate parameters. The current response change characteristics are processed based on the positioning accuracy drive source to generate a shockless start excitation current command. A fusion data network is established based on monitoring the operating status of multiple sensors using the non-impact start excitation current command. The configuration is reconstructed by activating the abnormal response signal in the fusion data network. A degraded operation fault-tolerant mode is generated based on the signal reconstruction configuration and the current response change characteristics. The shutdown command signal is received through the degraded operation fault tolerance mode, the shutdown command signal is converted into a response preparation gain, and the shutdown status is evaluated and an emergency shutdown trigger condition is generated based on the response preparation gain and the signal reconstruction configuration. The bus capacitor power supply status is detected based on the emergency stop triggering conditions. Gradient control potential is released from the energy changes of the bus capacitor power supply status. Based on the gradient control potential and the degraded operation fault tolerance mode, the active damping control timing is determined. The damping force control command is generated through the active damping control timing to realize intelligent start-stop control of the magnetic levitation refrigeration compressor.

2. The method according to claim 1, characterized in that, The step of performing fluctuation analysis on the current response data to obtain the location awareness enhancement source includes: Frequency anomaly detection is performed on current response data to obtain inductance field fluctuation characteristics; Identify the position error compensation current sequence from the inductor field fluctuation characteristics; The amplitude of the position error compensation current sequence is calibrated to form rotor positioning characteristics; A position awareness enhancement source is constructed using the rotor positioning features.

3. The method according to claim 1, characterized in that, The configuration reconstruction via an anomaly response activation signal in the fused data network includes: Detect sensor drift patterns using the abnormal response of the fused data network; Multi-channel collaborative calibration is performed using the sensor drift mode to form a collaborative calibration trajectory sequence; The coordinated calibration trajectory sequence is spatially reconstructed and mapped to obtain a set of sensor compensation coordinates; A signal reconstruction configuration is formed based on the sensor compensation coordinate set.

4. The method according to claim 1, characterized in that, The step of extracting the positioning accuracy drive source from the estimated deviation of the rotor offset coordinate parameters includes: The estimated deviation of the rotor offset coordinate parameters is analyzed by a sliding window to obtain the local deviation peak value; Gradient analysis is performed on the local deviation peaks arranged in time sequence to determine the deviation change trend; Identify deviation inflection features based on the slope abrupt change point of the deviation change trend; A positioning accuracy drive source is generated based on the time distribution density of the deviation turning point features.

5. The method according to claim 1, characterized in that, The step of converting the shutdown command signal into a response preparation gain includes: The shutdown command signal is decomposed into a dynamic delay sequence according to a time window; Buffer state analysis is performed on the dynamic delay sequence to generate a buffer time matrix; The dynamic delay sequence is phase-matched with the buffer time matrix to form a timing buffer diagram; The response preparation gain is determined by extracting the preparation enhancement nodes from the time-series buffer graph.

6. The method according to claim 1, characterized in that, Releasing gradient control potential from energy changes in the power supply state of the bus capacitor includes: Utilizing bus capacitor power supply status to monitor voltage attenuation gradient mode; An optimized discharge region is formed by performing capacitor discharge optimization control through the voltage decay gradient mode. The optimized discharge region is subjected to graded current regulation to obtain the current supply direction distribution; Gradient control potential is formed based on the aforementioned current supply direction distribution.

7. The method according to claim 1, characterized in that, The generation of damping force control commands through the active damping control timing includes: The torque oscillation pattern of the active damping control timing is monitored to generate a torque oscillation trajectory; Injecting an antiphase force at the peak position of the torque oscillation trajectory forms a rectangular torque standing wave; Deterministic torque components are extracted using the stable nodes of the torque standing wave; The deterministic torque components are superimposed in phase to generate a damping force control command.

8. The method according to claim 5, characterized in that, The step of phase matching the dynamic delay sequence with the buffer time matrix to form a timing buffer map includes: The dynamic delay sequence is converted into a virtual oscillation signal; The virtual oscillation signal is allowed to propagate in the buffer time matrix to generate an interference pattern; Extract the enhanced and reduced nodes from the interference pattern; The spatial distribution of the enhancement nodes and reduction nodes is encoded as a temporal buffer graph.

9. The method according to claim 7, characterized in that, The extraction of deterministic torque components using the stable nodes of the torque standing wave includes: Oscillation energy analysis is performed on the torque standing wave to obtain oscillation stability data; Identify torque optimization nodes and control optimization nodes in the oscillation stability data; Based on the torque optimization node and the control optimization node, parameter fusion is performed to generate a torque coordination curve; Feature extraction and analysis are performed on the torque coordination curve to form deterministic torque components.

10. An intelligent start-stop control system for a magnetic levitation refrigeration compressor, characterized in that, include: The data acquisition module is used to acquire current response data of the magnetic bearing coil, perform fluctuation analysis on the current response data to obtain the position sensing enhancement source, and obtain the current response change characteristics based on the position sensing enhancement source. The position estimation module is used to estimate the rotor position based on the current response change characteristics, extract rotor offset coordinate parameters, extract the positioning accuracy drive source from the estimation deviation of the rotor offset coordinate parameters, and process the current response change characteristics based on the positioning accuracy drive source to generate a shockless start excitation current command. The fault monitoring module is used to monitor the operating status of multiple sensors based on the non-impact start excitation current command, establish a fusion data network, reconstruct the configuration by activating the abnormal response signal in the fusion data network, and generate a degraded operation fault-tolerant mode based on the signal reconstruction configuration and the current response change characteristics. The shutdown control module is used to receive shutdown command signals through the degraded operation fault-tolerant mode, convert the shutdown command signals into response preparation gains, and perform shutdown state assessment based on the response preparation gains and the signal reconstruction configuration to generate emergency shutdown trigger conditions. The damping execution module is used to detect the power supply status of the bus capacitor according to the emergency stop triggering condition, release gradient control potential from the energy change of the power supply status of the bus capacitor, determine the active damping control timing based on the gradient control potential and the degraded operation fault tolerance mode, and generate a damping force control command through the active damping control timing to realize intelligent start-stop control of the magnetic levitation refrigeration compressor.

Images