Permanent magnet synchronous motor failure monitoring method, system and equipment for space environment

Abstract

The invention discloses a permanent magnet synchronous motor failure monitoring method, system and equipment for a space environment, and relates to the technical field of motor fault monitoring, and the method comprises the steps: constructing a simulation space environment experiment module, carrying out the accelerated aging experiment of a permanent magnet synchronous motor, and obtaining multi-dimensional degradation data; performing material performance evaluation on the permanent magnet synchronous motor, determining a material performance degradation result, performing failure analysis, and constructing a space environment failure feature library; and monitoring a motor operation parameter set of the permanent magnet synchronous motor in real time, determining an abnormal mode, and making a failure maintenance suggestion. The technical problems that in the prior art, the failure condition of the permanent magnet synchronous motor in the space environment cannot be accurately monitored, the reliability of the motor in the space environment is poor, and the operation stability is insufficient are solved, the failure condition of the permanent magnet synchronous motor is accurately monitored, and the reliability of the motor is improved. And the reliability and the operation stability of the permanent magnet synchronous motor in a space environment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of motor fault monitoring, and particularly to a method, system and device for monitoring the failure of a permanent magnet synchronous motor in a space environment. Background Art

[0002] In the aerospace field, the space environment is complex and harsh, with factors such as high vacuum, strong radiation, and extreme temperature changes. As a driving device for key components of satellites and other spacecraft, the reliable operation of a permanent magnet synchronous motor is crucial. However, existing motor failure monitoring technologies are difficult to meet the special requirements of the space environment, unable to comprehensively obtain the degradation data of the motor in this environment, and there are also biases in the evaluation of material properties under the influence of complex factors, resulting in difficulty in accurately judging the failure situation of the motor. Moreover, due to the lack of accurate failure monitoring and effective risk prediction, sudden failures often occur during the operation of the motor, seriously affecting the normal operation of the spacecraft.

[0003] The existing technology has technical problems such as being unable to accurately monitor the failure situation of a permanent magnet synchronous motor in a space environment, and the poor reliability and insufficient operation stability of the motor in a space environment. Summary of the Invention

[0004] This application provides a method, system and device for monitoring the failure of a permanent magnet synchronous motor in a space environment, aiming to solve the technical problems in the existing technology that it is impossible to accurately monitor the failure situation of a permanent magnet synchronous motor in a space environment, and the poor reliability and insufficient operation stability of the motor in a space environment.

[0005] In view of the above problems, this application provides a method, system and device for monitoring the failure of a permanent magnet synchronous motor in a space environment.

[0006] In the first aspect of this application, a method for monitoring the failure of a permanent magnet synchronous motor in a space environment is provided. The method includes: Construct a simulated space environment test chamber, which includes extreme temperature cycling conditions and high vacuum conditions. Conduct an accelerated aging test on the permanent magnet synchronous motor through the simulated space environment test chamber to obtain multi-dimensional degradation data; based on the multi-dimensional degradation data, evaluate the material properties of the permanent magnet synchronous motor, determine the material property degradation result, conduct failure analysis according to the material property degradation result, and construct a space environment failure feature library; monitor the motor operation parameter set of the permanent magnet synchronous motor in real time, match the motor operation parameter set with the space environment failure feature library to determine the abnormal mode, predict the failure risk level according to the abnormal mode, and formulate failure maintenance suggestions according to the failure risk level.

[0007] In a possible implementation, a simulated space environment test chamber is constructed. The simulated space environment test chamber includes extreme temperature cycling conditions and high vacuum conditions. An accelerated aging test is performed on a permanent magnet synchronous motor through the simulated space environment test chamber to obtain multi-dimensional degradation data, and the following processing is performed: Perform a stepped temperature cycle according to the extreme temperature cycling conditions of the simulated space environment test chamber, and arrange multi-type sensor arrays at the motor stator winding, the surface of the permanent magnet, and the bearing housing of the permanent magnet motor according to the high vacuum conditions of the simulated space environment test chamber to synchronously collect temperature distribution signals, vibration signals, and current harmonic signals; Filter and denoise the temperature distribution signals to obtain motor temperature distribution data, perform empirical mode decomposition on the vibration signals to obtain vibration spectrum data, and perform a transformation on the current harmonic signals to obtain current harmonic data; Perform time-frequency domain fusion processing on the motor temperature distribution data, the vibration spectrum data, and the current harmonic data to generate the multi-dimensional degradation data.

[0008] In a possible implementation, based on the multi-dimensional degradation data, a material performance evaluation is performed on the permanent magnet synchronous motor to determine the material performance degradation result, and the following processing is performed: Based on the motor temperature distribution data, determine the number of temperature cycles and perform a demagnetization curve test on the permanent magnet synchronous motor to obtain the demagnetization rate of the permanent magnet; Based on the vibration spectrum data, perform crystal structure evolution on the bearing lubricating film of the permanent magnet synchronous motor to generate a lubrication failure coefficient; Based on the current harmonic data, calculate the partial discharge inception voltage of the permanent magnet synchronous motor material and draw an insulation performance degradation curve; Couple the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve in multiple physical fields to construct the material performance degradation result.

[0009] In a possible implementation, failure analysis is performed according to the material performance degradation result to construct a space environment failure feature library, and the following processing is performed: Based on the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve, perform a simulation to determine the prior probabilities of multiple events; Based on the prior probabilities of the multiple events, construct a failure tree structure for the multiple events, perform a weight analysis on the failure tree structure of the multiple events to determine multiple weight coefficients; Based on the failure tree structure of the multiple events and the multiple weight coefficients, construct a multi-dimensional failure feature space; Perform clustering analysis according to the multi-dimensional failure feature space to determine multiple cluster centers, where the multiple cluster centers include multiple failure rules, and there is a corresponding relationship between the multiple cluster centers and the multiple failure rules; Extract high-density failure features according to the multiple failure rules to construct the space environment failure feature library.

[0010] In a possible implementation manner, the motor operation parameter set of the permanent magnet synchronous motor is monitored in real time, and the following processing is performed: The permanent magnet synchronous motor is collected in real time through a high-frequency current probe to obtain three-phase current harmonic component parameters; the permanent magnet synchronous motor is collected in real time through a distributed optical fiber temperature sensor to obtain winding temperature gradient parameters; the permanent magnet synchronous motor is collected in real time through a vibration sensor to obtain axial vibration acceleration parameters; wavelet packet transform is performed based on the three-phase current harmonic component parameters to extract characteristic energy values in a fixed frequency band, and the characteristic energy values include current harmonic distortion rates; fluctuation calculation is performed based on the winding temperature gradient parameters to determine the standard deviation of the temperature gradient; feature separation is performed based on the axial vibration acceleration parameters to determine the fretting wear characteristic frequency, and the fretting wear characteristic frequency includes vibration characteristic energy entropy; the current harmonic distortion rate, the standard deviation of the temperature gradient, and the vibration characteristic energy entropy are integrated to construct a composite operation monitoring vector, and the composite operation monitoring vector is added to the motor operation parameter set.

[0011] In a possible implementation manner, the motor operation parameter set is matched with the space environment failure feature library to determine an abnormal mode, and the following processing is performed: Feature dimensionality reduction is performed based on the composite operation monitoring vector to determine a composite dimensionality reduction vector; the composite dimensionality reduction vector is used as an index to traverse the space environment failure feature library for similarity matching to generate multiple matching degrees; when there is a matching degree exceeding a preset warning value among the multiple matching degrees, an online reconfiguration instruction is triggered, and fault analysis is performed through the online reconfiguration instruction to construct multiple abnormal labels; the motor operation parameter set is traced back according to the multiple abnormal labels for identification, and the abnormal mode is determined according to the identification result.

[0012] In a possible implementation manner, the failure risk level is predicted according to the abnormal mode, and the following processing is performed: The bearing wear value and the winding insulation deterioration value of the permanent magnet synchronous motor are extracted according to the abnormal mode; multivariate degradation failure analysis is performed based on the bearing wear value and the winding insulation deterioration value to determine the joint failure probability; a Markov chain is constructed for prediction according to the joint failure probability to construct a risk evolution trend, and failure impact analysis is performed based on the risk evolution trend to set the failure risk level.

[0013] In a possible implementation manner, failure maintenance suggestions are formulated according to the failure risk level, and the following processing is performed: A preset critical value is set based on the failure risk level, and the joint failure probability is compared and judged with the preset critical value; when the joint failure probability exceeds the preset critical value, a demagnetization compensation instruction is activated; multi-objective optimization is performed through the demagnetization compensation instruction to construct multi-level failure warning data, and maintenance control is performed based on the multi-level failure warning data to formulate the failure maintenance suggestions.

[0014] In the second aspect of the present application, a permanent magnet synchronous motor failure monitoring system for a space environment is provided. The system includes: A multi-dimensional degradation data acquisition module for constructing a simulated space environment test chamber, which includes extreme temperature cycling conditions and high vacuum conditions. The permanent magnet synchronous motor is subjected to an accelerated aging test through the simulated space environment test chamber to obtain multi-dimensional degradation data; a feature library construction module for evaluating the material properties of the permanent magnet synchronous motor based on the multi-dimensional degradation data, determining the material property degradation result, performing failure analysis according to the material property degradation result, and constructing a space environment failure feature library; a failure maintenance recommendation formulation module for real-time monitoring of the motor operation parameter set of the permanent magnet synchronous motor, matching the motor operation parameter set with the space environment failure feature library, determining the abnormal mode, predicting the failure risk level according to the abnormal mode, and formulating failure maintenance recommendations according to the failure risk level.

[0015] In the third aspect of the present application, an electronic device is provided. The electronic device includes: a processor; a memory for storing executable instructions of the processor; wherein, the processor is used to execute the permanent magnet synchronous motor failure monitoring method for a space environment provided by the present application.

[0016] One or more technical solutions provided in the present application have at least the following technical effects or advantages: Construct a simulated space environment test chamber, subject the permanent magnet synchronous motor to an accelerated aging test through the simulated space environment test chamber to obtain multi-dimensional degradation data; evaluate the material properties of the permanent magnet synchronous motor based on the multi-dimensional degradation data, determine the material property degradation result, and construct a space environment failure feature library; real-time monitor the motor operation parameter set of the permanent magnet synchronous motor, match it with the space environment failure feature library, determine the abnormal mode, predict the failure risk level according to the abnormal mode, and formulate failure maintenance recommendations according to the failure risk level. The technical effect of accurately monitoring the failure situation of the permanent magnet synchronous motor and improving the reliability and operation stability of the permanent magnet synchronous motor in the space environment is achieved. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic flowchart of the permanent magnet synchronous motor failure monitoring method for a space environment provided by an embodiment of the present application.

[0019] Figure 2 Schematic structural diagram of the permanent magnet synchronous motor failure monitoring system for the space environment provided by the embodiments of the present application.

[0020] Figure 3 Schematic structural diagram of an electronic device provided by the present application.

[0021] Explanation of reference numerals: multi-dimensional degradation data acquisition module 10, feature library construction module 20, failure maintenance recommendation formulation module 30, processor 21, memory 22, input device 23, output device 24. Specific embodiments

[0022] The present application provides a method, a system and a device for monitoring the failure of a permanent magnet synchronous motor in a space environment, aiming to solve the technical problems in the prior art that the failure of the permanent magnet synchronous motor in the space environment cannot be accurately monitored, and the reliability and operation stability of the motor in the space environment are poor.

[0023] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0024] Embodiment 1, as Figure 1 shown, the present application provides a method for monitoring the failure of a permanent magnet synchronous motor in a space environment, and the method includes: Step S100: Construct a simulated space environment test chamber, the simulated space environment test chamber includes extreme temperature cycle conditions and high vacuum conditions, and perform an accelerated aging test on the permanent magnet synchronous motor through the simulated space environment test chamber to obtain multi-dimensional degradation data.

[0025] Specifically, first, a simulated space environment test chamber with extreme temperature cycling conditions and high vacuum conditions is constructed to simulate the extreme temperature cycling conditions of the space environment. The stepped temperature cycling method is adopted to simulate the drastic temperature change process that the motor experiences in space. The high vacuum condition mimics the vacuum environment of space to reduce the interference of environmental factors on the experimental results. The permanent magnet synchronous motor is placed in this test chamber to carry out the accelerated aging experiment. During the experiment, in order to comprehensively capture various change data of the motor during the aging process, multi-type sensor arrays are arranged at key parts such as the stator winding of the motor, the surface of the permanent magnet, and the bearing seat to synchronously collect temperature distribution signals, vibration signals, and current harmonic signals. After these original signals are collected, the temperature distribution signal is filtered and denoised to remove the noise interference in the signal and obtain accurate motor temperature distribution data; for the vibration signal, the empirical mode decomposition technology is used to decompose the complex vibration signal into multiple intrinsic mode functions, and then vibration spectrum data is obtained to clearly present the vibration characteristics of different frequency components; specific transformation operations are performed on the current harmonic signal to obtain current harmonic data for analyzing the harmonic characteristics of the current. Finally, the processed motor temperature distribution data, vibration spectrum data, and current harmonic data are subjected to time-frequency domain fusion processing to integrate different types and different dimensional data together to generate multi-dimensional degradation data that can comprehensively reflect the performance degradation of the motor in the simulated space environment.

[0026] Step S200: Based on the multi-dimensional degradation data, evaluate the material properties of the permanent magnet synchronous motor, determine the material property degradation results, conduct failure analysis according to the material property degradation results, and construct a space environment failure feature library.

[0027] Specifically, based on the obtained multi-dimensional degradation data, in-depth analysis is carried out. Using the motor temperature distribution data, after determining the number of temperature cycles, the demagnetization curve test of the permanent magnet of the permanent magnet synchronous motor is carried out, and the demagnetization rate of the permanent magnet is obtained, which intuitively reflects the attenuation of the magnetism of the permanent magnet in the simulated space environment. Based on the vibration spectrum data, the crystal structure evolution of the bearing lubricating film of the permanent magnet synchronous motor is studied, and the lubrication failure coefficient is generated through quantitative analysis to measure the degradation degree of the bearing lubrication performance. With the help of the current harmonic data, the partial discharge inception voltage of the permanent magnet synchronous motor material is calculated, and the insulation performance degradation curve is drawn accordingly to clearly present the performance change trend of the insulation material. The demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve are coupled in multiple physical fields, comprehensively considering various factors, and the material performance degradation result is constructed to reveal the performance change of the motor material in the simulated space environment as a whole. Based on this material performance degradation result, the failure analysis is started. First, simulations are carried out according to the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve to determine the prior probabilities of multiple events related to motor failure. Then, the failure tree structure of multiple events is constructed using these prior probabilities, and the importance of each event in the failure tree is clarified through weight analysis to determine multiple weight coefficients. Combining the failure tree structure and the weight coefficients, a multi-dimensional failure feature space is constructed to provide a comprehensive mathematical model for analyzing the motor failure characteristics. Cluster analysis is carried out in the multi-dimensional failure feature space to determine multiple cluster centers, and each cluster center is associated with specific failure rules, which reflect the characteristic patterns of the motor under different failure conditions. Finally, high-density failure features are extracted according to the failure rules and integrated to construct a space environment failure feature library. This feature library covers the characteristic information of various failure situations of the permanent magnet synchronous motor in the simulated space environment, providing a key reference for subsequent real-time monitoring of the motor operating state, judging abnormal patterns, and predicting the failure risk level.

[0028] Step S300: Real-time monitor the set of motor operating parameters of the permanent magnet synchronous motor, match the set of motor operating parameters with the space environment failure feature library, determine the abnormal pattern, predict the failure risk level according to the abnormal pattern, and formulate failure maintenance suggestions according to the failure risk level.

[0029] Specifically, the motor operation parameter set of the permanent magnet synchronous motor is monitored in real time. These parameters cover key indicators such as the motor speed, current, voltage, temperature, etc. Subsequently, the obtained motor operation parameter set is carefully matched with the pre-constructed spatial environment failure feature library. This failure feature library stores various abnormal feature patterns that occur in the motor under different spatial environment conditions. Through comparison and analysis, the abnormal pattern presented by the current motor operation is accurately determined. Then, based on the determined abnormal pattern, its failure risk level is predicted. This level is usually divided into low, medium, and high degrees to intuitively reflect the likelihood of the motor experiencing a serious fault. Finally, according to the predicted failure risk level, targeted failure maintenance suggestions are formulated. For example, when the risk level is low, it can be recommended to increase the frequency of daily inspections; when the risk level is medium, relevant spare parts need to be prepared and technical personnel are arranged to closely monitor; when the risk level is high, the machine should be immediately shut down for a comprehensive overhaul and maintenance, so as to ensure the safe, stable, and efficient operation of the permanent magnet synchronous motor in a complex spatial environment.

[0030] In a possible implementation manner, step S100 further includes: Step S110: Perform a stepped temperature cycle according to the extreme temperature cycle condition of the simulated space environment realization chamber, and arrange multi-type sensor arrays at the motor stator winding, permanent magnet surface, and bearing seat of the permanent magnet motor according to the high vacuum condition of the simulated space environment realization chamber to synchronously collect temperature distribution signals, vibration signals, and current harmonic signals.

[0031] Step S120: Filter and denoise the temperature distribution signal to obtain motor temperature distribution data, perform empirical mode decomposition on the vibration signal to obtain vibration spectrum data, and perform transformation on the current harmonic signal to obtain current harmonic data.

[0032] Step S130: Perform time-frequency domain fusion processing on the motor temperature distribution data, the vibration spectrum data, and the current harmonic data to generate the multi-dimensional degradation data.

[0033] Specifically, the simulated space environment cabin sets extreme temperature cycle conditions and high vacuum conditions to highly restore the real space environment conditions. When performing the experiment, the ladder temperature cycle is strictly carried out in accordance with the extreme temperature cycle conditions of the simulated space environment cabin. This temperature change mode can quickly accelerate the aging of the motor, simulate the working state of the motor in the complex temperature environment of space, and let the motor experience a temperature change similar to that in space in a short time, so as to more efficiently obtain its performance change data under temperature stress. At the same time, according to the high vacuum conditions of the simulated space environment cabin, multi-type sensor arrays are arranged at key positions such as the motor stator winding, permanent magnet surface, and bearing seat of the permanent magnet motor. As a key component for the conversion of electrical energy and mechanical energy, the temperature and current changes of the motor stator winding can directly reflect the working state of the motor; the temperature change on the surface of the permanent magnet is closely related to the change of the magnetic field, which affects the magnetism and performance of the motor; the bearing seat is related to the mechanical operation stability of the motor, and its vibration is an important indicator for evaluating the mechanical performance of the motor. By deploying multiple types of sensors in these key locations, temperature distribution signals, vibration signals, and current harmonic signals can be collected simultaneously to comprehensively capture various aspects of real-time data when the motor is running in a simulated space environment, providing rich and critical original data support for subsequent in-depth analysis of motor performance degradation laws and construction of a failure monitoring system.

[0034] The temperature distribution signals collected from the permanent magnet motor under the simulated space environment are often mixed with noise due to factors such as actual environmental interference and the characteristics of the sensors themselves. These noises will seriously interfere with the accurate judgment of the true temperature of the motor. Therefore, filtering and noise reduction techniques are adopted. Through filtering algorithms, such as Kalman filtering or wavelet filtering, etc., the temperature distribution signals are processed to effectively remove the noise components, and then the motor temperature distribution data that can accurately reflect the actual temperature conditions of each part of the motor is obtained. For the collected vibration signals, the empirical mode decomposition (EMD) method is used for processing. The vibration signals contain rich information about the mechanical system of the motor, but they are usually complex non-linear and non-stationary signals. The empirical mode decomposition can decompose the vibration signals into multiple intrinsic mode functions (IMFs) according to the characteristic time scales of the signals themselves. These intrinsic mode functions represent the fluctuation components of the signals at different time scales. By analyzing the decomposed results, vibration spectrum data is obtained, from which the vibration energy distribution of different frequency components can be clearly understood, so as to judge whether there are abnormalities in the mechanical structure of the motor. For example, problems such as bearing wear and component loosening may be reflected in the vibration spectrum data. For the current harmonic signals, since their original forms are not conducive to directly analyzing the changes in the electrical performance of the motor, transformation processing is required. Usually, means such as Fourier transform or wavelet transform are used to convert the current harmonic signals from the time domain to the frequency domain, so as to obtain the current harmonic data. In the frequency domain, information such as the frequency components, amplitude sizes, and phase relationships of the current harmonics can be more intuitively observed. These current harmonic data are of great significance for evaluating the winding insulation performance, electromagnetic compatibility, etc. of the motor, and can help the monitoring personnel to detect potential fault hazards in the electrical system of the motor in time. The motor temperature distribution data, vibration spectrum data, and current harmonic data obtained after this series of processes lay the foundation for subsequent generation of multi-dimensional degradation data, construction of a space environment failure feature library, etc.

[0035] The wavelet transform algorithm is used to perform time-frequency domain fusion processing on the motor temperature distribution data, vibration spectrum data, and current harmonic data. First, wavelet decomposition is performed on the motor temperature distribution data to obtain the low-frequency approximation coefficients and high-frequency detail coefficients at different scales. The low-frequency approximation coefficients reflect the trend of temperature change, and the high-frequency detail coefficients contain the mutation information of the temperature. Similarly, wavelet decomposition is performed on the vibration spectrum data and current harmonic data to obtain their respective characteristic coefficients at different scales. Then, according to the physical meaning and correlation of the data, the decomposed coefficients are weighted and fused. For example, for the characteristic frequency bands that reflect the performance changes of the key components of the motor, the weights of their coefficients are increased. Finally, the fused coefficients are reconstructed through wavelet inverse transform to generate multi-dimensional degradation data containing the time-frequency characteristics of the motor temperature, vibration, and current harmonics, comprehensively presenting the degradation status of the motor under the simulated space environment.

[0036] In a possible implementation manner, step S200 further includes: Step S210: Based on the motor temperature distribution data, determine the number of temperature cycles to perform a demagnetization curve test on the permanent magnet synchronous motor to obtain the demagnetization rate of the permanent magnet.

[0037] Step S220: Based on the vibration spectrum data, perform crystal structure evolution on the bearing lubricating film of the permanent magnet synchronous motor to generate a lubrication failure coefficient.

[0038] Step S230: Based on the current harmonic data, calculate the partial discharge inception voltage of the materials of the permanent magnet synchronous motor and draw an insulation performance degradation curve.

[0039] Step S240: Couple the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve in a multi-physical field to construct the material performance degradation result.

[0040] Specifically, after obtaining the motor temperature distribution data, based on this data, determine the number of temperature cycles to perform a demagnetization curve test on the permanent magnet synchronous motor to obtain the demagnetization rate of the permanent magnet. First, identify the characteristics of the temperature cycle from the motor temperature distribution data, including the range, frequency, and amplitude of temperature changes, etc., to determine the number of temperature cycles. Then, according to the determined number of temperature cycles, perform a demagnetization curve test on the permanent magnet synchronous motor, and measure the magnetic performance parameters of the permanent magnet, such as remanence, coercivity, etc., before and after each temperature cycle. By analyzing the changes in these magnetic performance parameters during the temperature cycle, calculate the demagnetization rate of the permanent magnet. This demagnetization rate reflects the performance degradation degree of the permanent magnet under the action of temperature cycle and provides an important basis for evaluating the reliability of the permanent magnet synchronous motor in the space environment.

[0041] Based on the vibration spectrum data, the crystal structure evolution of the bearing lubricating film of a permanent magnet synchronous motor is analyzed and a lubrication failure coefficient is generated, which is achieved by the following machine learning algorithms: First, the vibration spectrum data is decomposed by wavelet packet decomposition, and the energy features of different frequency bands are extracted as input feature vectors; at the same time, the crystal structure parameters of the lubricating film (such as grain size, lattice distortion degree, etc.) are obtained through X-ray diffraction (XRD) analysis as output labels. A hybrid model that combines a bidirectional long short-term memory network (Bi-LSTM) and a convolutional neural network (CNN) is constructed: the CNN is used to extract the spatial features of the vibration spectrum and capture local frequency band anomalies; the Bi-LSTM is used to mine the temporal features and analyze the historical dependence of the crystal structure evolution. A transfer learning strategy is adopted to pre-train the model parameters with laboratory accelerated aging data and then fine-tune them with actual working condition data. The attention mechanism is introduced during model training to automatically assign weights to different frequency band features and highlight the features strongly related to lubrication failure (such as the energy change at the bearing passing frequency). Finally, the crystal structure parameters output by the model are input into a predefined physical degradation model, and a lubrication failure coefficient between 0 and 1 is generated through fuzzy logic reasoning to realize the quantitative evaluation of the bearing lubrication state.

[0042] After obtaining the current harmonic data of the permanent magnet synchronous motor, first perform a fast Fourier transform (FFT) on the current harmonic data to convert the current harmonic signal in the time domain to the frequency domain, accurately separate each harmonic component, and obtain the amplitude and phase information of each harmonic. Using these harmonic features, combined with parameters such as the dielectric constant and thickness of the motor insulation material, according to the electrical physical model of partial discharge, the Poisson equation and Maxwell's equations are solved by the finite element analysis method to calculate the electric field distribution inside the motor. Set the criterion for the initiation of partial discharge, such as the electric field strength reaching a certain threshold, to determine the partial discharge initiation voltage. As the motor operates, new current harmonic data is collected regularly, and the above analysis and calculation process is repeated, and the partial discharge initiation voltage values at different operating times are recorded. The polynomial fitting method is used to process these discrete voltage values, construct the functional relationship between the partial discharge initiation voltage and time, and then draw the insulation performance degradation curve to clearly show the degradation trend of the motor insulation performance over time.

[0043] After obtaining the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve, a multi-physics coupling method is used to construct the material property degradation results. First, a multi-field coupling model framework is established, integrating the electromagnetic-thermal-structural-fluid multi-physics through shared nodes and boundary conditions. For the demagnetization rate of the permanent magnet, the Jiles-Atherton model is used to describe the change of the hysteresis loop, and a temperature correction coefficient is introduced to map the temperature distribution data in the thermal field into the electromagnetic field calculation. For the lubrication failure coefficient, the hydrodynamic characteristics of the bearing lubricating film are simulated by computational fluid dynamics (CFD), and the influence of the crystal structure evolution of the lubricating film on the viscosity and friction coefficient is combined with molecular dynamics simulation to convert the mechanical vibration data into fluid boundary conditions. In terms of insulation performance, based on the partial discharge inception voltage data, the finite element method is used to solve the electric field distribution, the Weibull statistical model is combined to describe the insulation aging process, and the heat generated by partial discharge is fed back into the thermal field calculation through the heat conduction equation. The coupling algorithm adopts a sequential iterative solution strategy. First, the magnetic field change caused by the demagnetization of the permanent magnet is calculated in the electromagnetic field and transferred to the structural field as a load; after the structural field calculates the bearing vibration response, the vibration parameters are transferred to the fluid field to correct the lubricating film characteristics; data exchange is carried out between the fluid field and the thermal field through the convective heat transfer coefficient; finally, the thermal field temperature distribution is fed back into the electromagnetic field and the insulation aging model. By setting reasonable convergence criteria (such as the energy error is less than 5%), multiple rounds of iteration are carried out until a stable solution is obtained. The final output includes the comprehensive material property degradation curve of magnetic property degradation, mechanical wear, and insulation aging, as well as the life prediction results of key components, providing a quantitative basis for the reliability assessment of permanent magnet synchronous motors in the space environment.

[0044] In a possible implementation manner, step S200 further includes: Step S250: Based on the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve, perform a simulation to determine the prior probabilities of multiple events.

[0045] Step S260: Based on the prior probabilities of multiple events, construct a fault tree structure for multiple events, and perform a weight analysis on the fault tree structure of multiple events to determine multiple weight coefficients.

[0046] Step S270: Based on the fault tree structure of multiple events and the multiple weight coefficients, construct a multi-dimensional failure feature space.

[0047] Step S280: Perform a clustering analysis according to the multi-dimensional failure feature space to determine multiple cluster centers, where the multiple cluster centers include multiple failure rules, and there is a corresponding relationship between the multiple cluster centers and the multiple failure rules.

[0048] Step S290: Extract high-density failure features according to the multiple failure rules, and construct the space environment failure feature library.

[0049] Specifically, to determine the prior probabilities of multiple events, first, data such as the permanent magnet demagnetization rate, lubrication failure coefficient, and insulation performance degradation curve are sorted and quantified. Using finite element analysis software, a multi-physics field coupling simulation model including the motor magnetic circuit, mechanical structure, and insulation system is constructed. In the model, parameter variables related to permanent magnet demagnetization, bearing lubrication failure, and insulation performance degradation are set respectively. The permanent magnet demagnetization rate is converted into a change in magnetic permeability and input into the magnetic circuit module. The lubrication failure coefficient is reflected in the changes in the bearing friction coefficient and lubrication film thickness. The insulation performance degradation curve is used to set the change in the dielectric strength of the insulation material over time. Combining the historical data of the motor operation in the space environment, the operation process of the motor under different working conditions is simulated. In each simulation run, the number of occurrences of key failure events such as the permanent magnet demagnetization reaching a certain degree, the bearing lubrication film completely failing, and insulation breakdown is counted. After a large number of simulation experiments, the number of occurrences of each failure event is divided by the total number of simulation runs to obtain the frequency of each failure event occurring under the current conditions, and this frequency is approximately used as the prior probability of multiple events.

[0050] Taking the motor system failure as the top event, the events related to permanent magnet demagnetization, lubrication failure, and insulation performance degradation are used as intermediate events and bottom events, and a failure tree is constructed by connecting them according to the logical causal relationship. For example, if the permanent magnet demagnetization causes the motor torque to decrease, which in turn affects the system operation, the permanent magnet demagnetization event is taken as the bottom event of the event causing the motor torque to decrease. The analytic hierarchy process (AHP) is used for weight analysis to construct a judgment matrix. First, the factors affecting the importance of failure events are determined, such as the event occurrence probability, the degree of influence on the motor performance, and the maintenance difficulty. For each factor, the relative importance of different failure events is compared pairwise and assigned values to construct the judgment matrix. Then, the maximum eigenvalue and eigenvector of the judgment matrix are calculated, and the rationality of the judgment is ensured through a consistency test. The elements in the eigenvector are the relative weights of each failure event corresponding to this factor. Considering the weights under all factors comprehensively, the weighted average method is used to obtain the final weight coefficient of each failure event.

[0051] Construct a four-dimensional failure feature space based on the fault tree structure and weight coefficients. The specific implementation is as follows: Map the bottom events of the fault tree to the four dimensions of the feature space. Among them, the demagnetization rate η of the permanent magnet is used as the magnetic performance dimension, the lubrication failure coefficient is converted into the mechanical degradation degree δ as the mechanical performance dimension, the bearing wear amount is converted into the wear degree ω as the structural performance dimension, and the natural logarithm of the space environment vacuum degree P is used as the environmental dimension. Use the prior probability of the bottom event as the coordinate origin of each dimension, and the weight coefficient as the scaling factor to weight the coordinates. For the logic gates in the fault tree, calculate the position vectors of the intermediate events and the top event in the four-dimensional space through the probability propagation algorithm. For example, if an intermediate event is connected by the demagnetization rate η and the degradation degree δ through an AND gate, its coordinate value is the product of the coordinate values of the two bottom events multiplied by the corresponding weight coefficient. Introduce the time dimension to form a dynamic coordinate system, and use parameters that change with time, such as the insulation performance degradation curve, as the dynamic coordinates on the time axis. Use Monte Carlo simulation to generate a large number of random sample points to verify the completeness of the feature space. Finally, construct a failure feature space containing four dimensions of magneto-mechanical-structure-environment, and each dimension carries physical meaning and weight information, providing a structured data basis for subsequent clustering analysis.

[0052] After completing the construction of the multi-dimensional failure feature space, use the clustering analysis algorithm to process the data points in this space. The clustering analysis will divide them into different clusters according to the similarity between the data points, and each cluster has a corresponding clustering center. During the analysis process, clustering algorithms such as K-Means are used, and through continuous iterative calculations, the data points gradually approach their most suitable clustering centers. While determining the clustering centers, deeply mine the characteristics of the data points covered by each clustering center. According to the distribution laws of the dimension data such as the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve, summarize the failure rules applicable to this clustering center. For example, if most of the data points under a certain clustering center show the characteristics that the demagnetization rate of the permanent magnet exceeds a certain threshold and the lubrication failure coefficient is also in a specific interval, then the demagnetization rate of the permanent magnet greater than X and the lubrication failure coefficient in the Y-Z interval can be set as a failure rule. Each clustering center corresponds to a unique set of failure rules, and these rules can accurately describe the motor failure mode represented by this clustering center, providing a key basis for subsequent construction of the space environment failure feature library.

[0053] After obtaining multiple failure rules, comprehensively sort out and analyze the data involved in these rules. For each failure rule, select data samples that meet the rule from a large amount of motor operation data. Various failure characteristics included in these samples, such as the permanent magnet demagnetization rate, lubrication failure coefficient, insulation performance degradation-related parameters, etc., are the objects that need to be focused on. Through statistical analysis methods, calculate the occurrence frequency and distribution of each characteristic in the data samples that meet the corresponding failure rules. Determine the high-density failure characteristics for those characteristics with high occurrence frequencies and key indication roles in failure determination. For example, if under a certain type of failure rule, when the permanent magnet demagnetization rate reaches a certain value range, the probability of the motor failure is extremely high, then this demagnetization rate value range belongs to the high-density failure characteristics. Integrate and classify all the high-density failure characteristics extracted from different failure rules, and organize them according to a certain storage structure and indexing method to build a space environment failure characteristic library. This characteristic library can quickly and accurately provide a reference for the abnormal mode judgment when real-time monitoring the operation parameters of the permanent magnet synchronous motor, greatly improving the efficiency and accuracy of motor failure monitoring.

[0054] In a possible implementation manner, step S300 further includes: Step S310: Perform real-time acquisition on the permanent magnet synchronous motor through a high-frequency current probe to obtain three-phase current harmonic component parameters.

[0055] Step S320: Perform real-time acquisition on the permanent magnet synchronous motor through a distributed optical fiber temperature sensor to obtain winding temperature gradient parameters.

[0056] Step S330: Perform real-time acquisition on the permanent magnet synchronous motor through a vibration sensor to obtain axial vibration acceleration parameters.

[0057] Step S340: Perform wavelet packet transform based on the three-phase current harmonic component parameters to extract the characteristic energy values of a fixed frequency band, and the characteristic energy values include the current harmonic distortion rate.

[0058] Step S350: Perform fluctuation calculation based on the winding temperature gradient parameters to determine the temperature gradient standard deviation.

[0059] Step S360: Perform feature separation based on the axial vibration acceleration parameters to determine the fretting wear characteristic frequency, and the fretting wear characteristic frequency includes the vibration characteristic energy entropy.

[0060] Step S370: Integrate the current harmonic distortion rate, the temperature gradient standard deviation, and the vibration characteristic energy entropy to construct a composite operation monitoring vector, and add the composite operation monitoring vector to the motor operation parameter set.

[0061] Specifically, parameter acquisition is carried out first. The three-phase current of the motor is collected in real time through a high-frequency current probe, and the time-domain current signal is converted into the frequency domain by using the fast Fourier transform (FFT) to obtain the three-phase current harmonic component parameters, which reflect the changes in the electrical performance of the motor.

[0062] At the same time, the temperature of the motor winding is monitored in real time with the help of a distributed fiber optic temperature sensor. This sensor can accurately measure the temperature at different positions of the winding, and then obtain the winding temperature gradient parameter, which can effectively reflect the thermal distribution inside the winding.

[0063] In addition, a vibration sensor is used to collect the axial vibration of the motor in real time to obtain the axial vibration acceleration parameter, which is closely related to the mechanical structure state of the motor.

[0064] Next, feature extraction and calculation are carried out. For the three-phase current harmonic component parameters collected, wavelet packet transform is used for processing. Wavelet packet transform can decompose the signal into different frequency bands. By extracting the characteristic energy values of a fixed frequency band, the current harmonic distortion rate can be calculated. This index is an important parameter for measuring the current quality of the motor.

[0065] After obtaining the winding temperature gradient parameter, fluctuation calculation is carried out on it to determine the temperature gradient standard deviation. First, the collected winding temperature gradient data are arranged in chronological order, and the difference in temperature gradient between adjacent time points is calculated to obtain a series of gradient change values. Then, the mean value of these gradient change values is calculated by using statistical methods. Next, by calculating the sum of the squares of the differences between each gradient change value and the mean value, and dividing by the number of data points minus 1 and then taking the square root, the temperature gradient standard deviation is obtained, which can effectively reflect the fluctuation degree of the winding temperature gradient.

[0066] For the axial vibration acceleration parameter, methods such as wavelet transform or time-frequency analysis are used for feature separation. Through these methods, the frequency components related to fretting wear are screened out from the complex vibration signal, and then the fretting wear characteristic frequency is determined. In this process, the energy distribution of the vibration signal in different frequency bands is calculated to obtain the vibration characteristic energy entropy, which can measure the complexity and disorder of the vibration signal.

[0067] Finally, the three key parameters of the current harmonic distortion rate, temperature gradient standard deviation, and vibration characteristic energy entropy obtained above are integrated. They are combined together in a specific order to construct a composite operation monitoring vector. For example, taking the current harmonic distortion rate as the first element, the temperature gradient standard deviation as the second element, and the vibration characteristic energy entropy as the third element to form a three-dimensional vector. After completion of the construction, this composite operation monitoring vector is added to the motor operation parameter set, providing an important basis for comprehensively evaluating the operation state of the permanent magnet synchronous motor in the future and facilitating the timely discovery of potential fault hazards of the motor.

[0068] In a possible implementation manner, step S300 further includes: Step S380: Perform feature dimensionality reduction based on the composite operation monitoring vector to determine a composite dimensionality reduction vector; use the composite dimensionality reduction vector as an index to traverse the space environment failure feature library for similarity matching to generate multiple matching degrees; when there is a matching degree exceeding a preset warning value among the multiple matching degrees, trigger an online reconfiguration instruction, perform fault analysis through the online reconfiguration instruction, construct multiple anomaly labels; trace back to the motor operation parameter set according to the multiple anomaly labels for identification, and determine the anomaly mode according to the identification result.

[0069] Specifically, after obtaining the composite operation monitoring vector, a method combining principal component analysis (PCA) and kernel principal component analysis (KPCA) is used for feature dimensionality reduction. First, the original vector is preprocessed by standardization to eliminate the influence of different parameter dimensions. By calculating the eigenvalues and eigenvectors of the feature covariance matrix, the principal components with a cumulative contribution rate exceeding a preset threshold are extracted to achieve linear dimensionality reduction. At the same time, KPCA with a Gaussian kernel function is applied to process non-linear features, and the data is mapped to a high-dimensional space through the kernel trick and then the principal components are extracted. The hypothesis testing method is used to analyze the difference between the two dimensionality reduction results, and the dimensionality reduction scheme is dynamically selected or fused according to the test results. Finally, through feature importance evaluation, the core dimensions with the highest sensitivity to faults are retained to construct a composite dimensionality reduction vector.

[0070] When using the composite dimensionality reduction vector as an index for similarity matching, first, the hash algorithm is used to preprocess the data in the space environment failure feature library. A unique hash value is generated for each failure feature vector, and the vector is stored in the corresponding hash bucket according to the hash value to construct a hash table. In this way, during subsequent searches, the area where the vector that may be matched can be quickly located, greatly reducing the traversal range. When calculating the similarity, the cosine similarity formula is used to measure the similarity between the composite dimensionality reduction vector and the vectors in the failure feature library. For each candidate vector taken out from the hash bucket, its cosine similarity with the composite dimensionality reduction vector is calculated. After traversing all relevant candidate vectors, a corresponding matching degree value is generated for each candidate vector. These matching degree values reflect the similarity between the composite dimensionality reduction vector and each candidate vector. The closer the value is to 1, the higher the similarity between the two vectors, and the closer the value is to 0, the lower the similarity. Finally, the generated multiple matching degrees are sorted into a list form for subsequent screening and analysis, providing a basis for determining the matching situation between the current operating state of the motor and the known failure modes.

[0071] When the monitoring system detects a match degree exceeding the preset warning value among multiple match degrees, it immediately triggers an online reconfiguration instruction. Start the fault analysis process based on deep learning, and use a convolutional neural network (CNN) to deeply mine fault features. First, take the composite dimensionality reduction vector as the input. Through multiple convolutional layers and pooling layers, the CNN automatically extracts the high-level fault features hidden in the data. These features are associated with the data patterns in the space environment failure feature library. Through the contrast learning algorithm, the understanding and discrimination ability of fault features are further enhanced. At the same time, combined with the classification algorithm based on decision trees, classify and make decisions on the extracted features. According to the preset decision rules, the decision tree gradually determines the fault type based on key parameters such as the current harmonic distortion rate, the standard deviation of the temperature gradient, and the vibration feature energy entropy. During the analysis process, use a support vector machine (SVM) to divide the boundary of the fault data to improve the accuracy and stability of classification. Finally, comprehensively combine the results of these algorithms to construct multiple anomaly labels. These labels cover various fault situations that may occur in the permanent magnet synchronous motor in the space environment, such as permanent magnet demagnetization, bearing wear, winding insulation aging, etc., providing a detailed and accurate basis for determining the abnormal mode of the motor in the follow-up.

[0072] After constructing multiple anomaly labels, based on the fault feature keywords in the labels, quickly locate to the original measurement data segment corresponding to the motor operation parameter set through spatio-temporal indexing. Extract parameter sequences such as the current harmonic distortion rate, the standard deviation of the temperature gradient, and the vibration feature energy entropy before and after the label generation moment, and apply Granger causality test to identify the strongly correlated parameter subset. Combine with a Bayesian network to construct a fault propagation probability model, analyze the temporal dependence and causal relationship strength of parameter changes. Use the dynamic time warping (DTW) algorithm to match the parameter sequence with the standard fault mode in the feature library. When the parameter subsets corresponding to multiple labels form a closed-loop causal chain and the similarity exceeds the preset threshold, it is confirmed as a specific abnormal mode. Finally, map the abnormal mode to the fault ontology library to generate a multi-dimensional diagnostic report including the fault type, development stage, and influence range, providing a decision-making basis for fault isolation and repair.

[0073] In a possible implementation manner, step S300 further includes: Step S390: Extract the bearing wear value and the winding insulation deterioration value of the permanent magnet synchronous motor according to the abnormal mode; conduct a multivariate degradation failure analysis based on the bearing wear value and the winding insulation deterioration value to determine the joint failure probability; construct a Markov chain for prediction according to the joint failure probability, construct a risk evolution trend, conduct a failure impact analysis based on the risk evolution trend, and set the failure risk level.

[0074] Specifically, after clarifying the abnormal modes of the permanent magnet synchronous motor, relevant key data extraction begins. For the bearing wear value, by means of the axial vibration acceleration parameters continuously collected by the vibration sensor, the time-domain vibration signal is converted into a frequency-domain signal through fast Fourier transform to obtain the vibration spectrum. Analyze the amplitude changes in specific frequency bands in the spectrum (such as the characteristic frequencies related to bearing faults), and combine with the pre-constructed mapping relationship model between the vibration amplitude and the bearing wear amount, which is obtained by fitting a large amount of experimental data. For example, when the vibration amplitude reaches a certain threshold in the frequency band of 100Hz - 200Hz, the corresponding bearing wear value can be deduced according to the model. For the winding insulation degradation value, comprehensive analysis is carried out by using the winding temperature gradient parameters obtained by the distributed optical fiber temperature sensor and the three-phase current harmonic component parameters collected by the high-frequency current probe. First, calculate the partial discharge inception voltage according to the current harmonic data, because partial discharge is an important manifestation of winding insulation degradation. Then, based on the correlation between the temperature gradient and the insulation aging rate, combined with the pre-drawn insulation performance degradation curve (obtained from aging experiments under different temperature and voltage conditions), determine the degree of winding insulation degradation over time at the current temperature and partial discharge inception voltage, and then obtain the winding insulation degradation value.

[0075] After obtaining the bearing wear value and the winding insulation degradation value, the long short-term memory network (LSTM) in deep learning is combined with the extreme learning machine (ELM) to determine the joint failure probability. First, normalize the bearing wear value, the winding insulation degradation value, and relevant data such as the ambient temperature and load change rate during the operation of the motor, and input the processed data into the LSTM network according to the time series. The LSTM network can effectively capture the long-term dependence relationships in the data, and through its internal forget gate, input gate, and output gate mechanisms, screen out the historical data features that have an important impact on motor failure. Then, input the feature vector output by the LSTM network into the extreme learning machine ELM. ELM randomly generates the connection weights between the input layer and the hidden layer and the thresholds of the hidden layer neurons. By only setting the number of hidden layer neurons, the training can be completed quickly. It can further classify and regress the features extracted by LSTM, and establish a complex non-linear relationship between the input features and motor failure. Finally, the result output by ELM is the joint failure probability of the motor.

[0076] Based on the determined combined failure probability, a Markov chain is constructed to predict the change of the motor state, the risk evolution trend is constructed, and the failure risk level is set. First, the operating states of the permanent magnet synchronous motor are divided into discrete states such as normal, slightly abnormal, moderately abnormal, and severe fault. According to the combined failure probability and the historical operating data of the motor, the transition probabilities between different states are determined, and a state transition probability matrix is constructed. For example, through data analysis, it is found that when the motor is in the normal state at the current moment, the probability of transitioning to the slightly abnormal state at the next moment is 0.1, and the probabilities of transitioning to other states are also determined accordingly and filled into the matrix. Then, using the constructed Markov chain, the future operating state changes of the motor are simulated through multiple iterations. In each iteration, the next state is randomly determined according to the current state and the state transition probability matrix. After multiple iterations, a series of state change sequences are obtained, and then the risk evolution trend is constructed to visually present the probability changes of the motor being in different states at different time points. Based on the risk evolution trend, the failure impact analysis is carried out to evaluate the impact degree of the motor on the entire system (such as the satellite system) under different risk states. For the slightly abnormal state of the motor, analyze its impact on aspects such as satellite attitude control accuracy and energy consumption; for the moderately abnormal and severe fault states, evaluate the impact on satellite mission execution and component damage respectively, and quantify these impacts, such as a 2% decrease in attitude control accuracy caused by slight abnormality and a 10% increase in energy consumption caused by moderate abnormality. Finally, the failure risk level is set according to the severity of the failure impact. The risk level is divided into three levels: low, medium, and high. If the motor is likely to maintain the normal state in the future for a period of time, and even if a slight abnormality occurs, the impact on the system is small, it is set as the low risk level; if there is a certain probability of entering the moderately abnormal state and it has an obvious impact on the system performance, it is set as the medium risk level; if there is a high probability of entering the severe fault state and it causes a serious impact on the system or even leads to mission failure, it is set as the high risk level. In this way, a clear basis is provided for subsequent targeted maintenance measures.

[0077] In a possible implementation manner, step S390 further includes: Step S391: Set a preset critical value based on the failure risk level, and compare and determine the combined failure probability with the preset critical value.

[0078] Step S392: When the combined failure probability exceeds the preset critical value, activate the demagnetization compensation instruction.

[0079] Step S393: Perform multi-objective optimization through the demagnetization compensation instruction, construct multi-level failure warning data, perform maintenance control based on the multi-level failure warning data, and formulate the failure maintenance suggestions.

[0080] Specifically, based on the set failure risk levels, corresponding preset critical values are set for the three levels of low, medium, and high. For example, the critical value corresponding to the low risk level is 0.3, the medium risk level is 0.6, and the high risk level is 0.8. The calculated combined failure probability is compared with these preset critical values for determination, so as to clarify the risk state of the current motor.

[0081] In the motor failure monitoring process, once the combined failure probability is calculated, it will be compared with the preset critical value in real time. Once the system determines that the combined failure probability exceeds the preset critical value, the demagnetization compensation mechanism will be immediately triggered and the demagnetization compensation instruction will be activated. At this time, the monitoring system quickly sends a signal to the control system of the motor. After receiving the instruction, the control system immediately starts the preset demagnetization compensation program. This program will accurately analyze the demagnetization situation of the motor based on the current operating parameters of the motor, such as the harmonic components of the three-phase current, the winding temperature gradient, and the axial vibration acceleration, combined with the data in the failure feature library accumulated in the early stage. Subsequently, the control system adjusts the control strategy of the motor according to the analysis results, such as changing the magnitude, phase, and frequency of the current, and generates a compensation magnetic field opposite to the direction of the demagnetization magnetic field through precise regulation of the current, so as to offset the weakening of the magnetic field caused by demagnetization, and then maintain the normal performance of the motor and ensure that the motor can still operate continuously and stably under the demagnetization risk.

[0082] Multi-objective optimization is carried out using the Particle Swarm Optimization (PSO) algorithm, with the improvement of motor efficiency, minimization of torque ripple, and control of temperature rise within a reasonable range as the optimization objectives. A group of particles is initialized, where each particle represents a set of motor control parameters (such as current amplitude, phase angle, etc.), and the positions and velocities of the particles are randomly distributed in the solution space. In each iteration, the particles update their velocities and positions based on their own historical best positions and the group's historical best positions, continuously approaching the optimal solution. Through multiple iterations of optimization, a series of compensation parameter combinations that meet the multi-objective requirements are obtained. Based on these optimized parameters, multi-level failure warning data is constructed in combination with the real-time operating data of the motor. The clustering algorithm (such as DBSCAN) is used to perform clustering analysis on the motor operating parameters, and the motor operating status is divided into four levels: normal, slightly abnormal, moderately abnormal, and severely abnormal. Corresponding thresholds are set for each level. For example, the current harmonic distortion rate is within 5% in the normal state, 5% - 10% in the slightly abnormal state, 10% - 20% in the moderately abnormal state, and greater than 20% in the severely abnormal state. Maintenance control is carried out based on the multi-level failure warning data, and failure maintenance suggestions are formulated. If it is in the normal state, regular periodic maintenance inspections are arranged; in the slightly abnormal state, the monitoring frequency is increased, and the changes in the motor operating parameters are closely monitored; in the moderately abnormal state, it is prompted to prepare for replacing vulnerable parts (such as bearings), and technicians are arranged to standby; in the severely abnormal state, an emergency alarm is immediately issued, the motor operation is stopped, and technicians are guided to conduct a comprehensive overhaul of the motor, such as replacing permanent magnets and repairing winding insulation, to ensure the safe and stable operation of the motor.

[0083] Embodiment 2, based on the same inventive concept as the permanent magnet synchronous motor failure monitoring method for space environment in the foregoing embodiment, as Figure 2 shown, the present application provides a permanent magnet synchronous motor failure monitoring system for space environment. The system in the embodiment of the present application and the method embodiment are based on the same inventive concept. Among them, the system includes: A multi-dimensional degradation data acquisition module 10, which is used to construct a simulated space environment test chamber. The simulated space environment test chamber includes extreme temperature cycle conditions and high vacuum conditions. Through the simulated space environment test chamber, an accelerated aging experiment is carried out on the permanent magnet synchronous motor to obtain multi-dimensional degradation data.

[0084] A feature library construction module 20, which is used to evaluate the material properties of the permanent magnet synchronous motor based on the multi-dimensional degradation data, determine the material property degradation results, perform failure analysis according to the material property degradation results, and construct a space environment failure feature library.

[0085] The failure maintenance recommendation formulation module 30 is used to monitor the motor operation parameter set of the permanent magnet synchronous motor in real time, match the motor operation parameter set with the space environment failure feature library, determine the abnormal mode, predict the failure risk level according to the abnormal mode, and formulate failure maintenance recommendations according to the failure risk level.

[0086] Furthermore, the system is also used to implement the following functions: Execute a stepped temperature cycle according to the extreme temperature cycle condition of the cabin that simulates the space environment, and deploy multi-type sensor arrays at the motor stator winding, the surface of the permanent magnet, and the bearing housing of the permanent magnet motor according to the high vacuum condition of the cabin that simulates the space environment to synchronously collect temperature distribution signals, vibration signals, and current harmonic signals; filter and denoise the temperature distribution signals to obtain motor temperature distribution data, perform empirical mode decomposition on the vibration signals to obtain vibration spectrum data, and perform transformation on the current harmonic signals to obtain current harmonic data; perform time-frequency domain fusion processing on the motor temperature distribution data, the vibration spectrum data, and the current harmonic data to generate the multi-dimensional degradation data.

[0087] Furthermore, the system is also used to implement the following functions: Based on the motor temperature distribution data, determine the number of temperature cycles to perform a demagnetization curve test on the permanent magnet synchronous motor to obtain the demagnetization rate of the permanent magnet; based on the vibration spectrum data, perform crystal structure evolution on the bearing lubricating film of the permanent magnet synchronous motor to generate a lubrication failure coefficient; based on the current harmonic data, calculate the partial discharge inception voltage of the materials of the permanent magnet synchronous motor and draw an insulation performance degradation curve; couple the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve in multiple physical fields to construct the material performance degradation result.

[0088] Furthermore, the system is also used to implement the following functions: Based on the demagnetization rate of the permanent magnet, the lubrication failure coefficient, and the insulation performance degradation curve, perform simulation to determine the prior probabilities of multiple events; construct a failure tree structure of multiple events based on the prior probabilities of multiple events, perform weight analysis on the failure tree structure of multiple events to determine multiple weight coefficients; construct a multi-dimensional failure feature space based on the failure tree structure of multiple events and the multiple weight coefficients; perform clustering analysis according to the multi-dimensional failure feature space to determine multiple cluster centers, and the multiple cluster centers include multiple failure rules, where there is a corresponding relationship between the multiple cluster centers and the multiple failure rules; extract high-density failure features according to the multiple failure rules to construct the space environment failure feature library.

[0089] Furthermore, the system is also used to implement the following functions: Real-time acquisition is performed on the permanent magnet synchronous motor through a high-frequency current probe to obtain three-phase current harmonic component parameters; real-time acquisition is performed on the permanent magnet synchronous motor through a distributed optical fiber temperature sensor to obtain winding temperature gradient parameters; real-time acquisition is performed on the permanent magnet synchronous motor through a vibration sensor to obtain axial vibration acceleration parameters; wavelet packet transform is performed based on the three-phase current harmonic component parameters to extract characteristic energy values in a fixed frequency band, and the characteristic energy values include current harmonic distortion rates; fluctuation calculation is performed based on the winding temperature gradient parameters to determine the standard deviation of the temperature gradient; feature separation is performed based on the axial vibration acceleration parameters to determine the fretting wear characteristic frequency, and the fretting wear characteristic frequency includes vibration characteristic energy entropy; the current harmonic distortion rate, the standard deviation of the temperature gradient, and the vibration characteristic energy entropy are integrated to construct a composite operation monitoring vector, and the composite operation monitoring vector is added to the motor operation parameter set.

[0090] Further, the system is also used to implement the following functions: Feature dimension reduction is performed based on the composite operation monitoring vector to determine a composite dimension reduction vector; using the composite dimension reduction vector as an index, traversing the spatial environment failure feature library for similarity matching to generate multiple matching degrees; when there is a matching degree exceeding a preset warning value among the multiple matching degrees, triggering an online reconfiguration instruction, performing fault analysis through the online reconfiguration instruction, and constructing multiple abnormal labels; backtracking to the motor operation parameter set according to the multiple abnormal labels for identification, and determining the abnormal mode based on the identification result.

[0091] Further, the system is also used to implement the following functions: Extract the bearing wear value and winding insulation deterioration value of the permanent magnet synchronous motor according to the abnormal mode; perform multivariate degradation failure analysis based on the bearing wear value and the winding insulation deterioration value to determine the joint failure probability; construct a Markov chain for prediction according to the joint failure probability, construct a risk evolution trend, perform failure impact analysis based on the risk evolution trend, and set a failure risk level.

[0092] Further, the system is also used to implement the following functions: Set a preset critical value based on the failure risk level, and compare and determine the joint failure probability with the preset critical value; when the joint failure probability exceeds the preset critical value, activate a demagnetization compensation instruction; perform multi-objective optimization through the demagnetization compensation instruction, construct multi-level failure warning data, perform maintenance control based on the multi-level failure warning data, and formulate the failure maintenance suggestion.

[0093] Embodiment 3 Figure 3Schematic diagram of the structure of the electronic device provided for the permanent magnet synchronous motor failure monitoring method for the space environment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing the embodiments of the present invention. Figure 3 The displayed electronic device is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present invention. As Figure 3 shown, the electronic device includes a processor 21, a memory 22, an input device 23, and an output device 24; the number of processors 21 in the electronic device can be one or more, Figure 3 taking one processor 21 as an example, the processor 21, the memory 22, the input device 23, and the output device 24 in the electronic device can be connected by a bus or other means, Figure 3 taking the connection by bus as an example.

[0094] It should be noted that the above-mentioned order of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above description of specific embodiments of this specification is made. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0095] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0096] This specification and the drawings are only exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for monitoring the failure of a permanent magnet synchronous motor used in a space environment, characterized in that, The method includes: Construct a simulated space environment test chamber, which includes extreme temperature cycling conditions and high vacuum conditions. Accelerated aging tests are performed on the permanent magnet synchronous motor through the simulated space environment test chamber to obtain multi-dimensional degradation data; Based on the multi-dimensional degradation data, perform material property evaluation on the permanent magnet synchronous motor, determine the material property degradation results, conduct failure analysis according to the material property degradation results, and construct a space environment failure feature library; Real-time monitor the set of motor operating parameters of the permanent magnet synchronous motor, match the set of motor operating parameters with the space environment failure feature library, determine the abnormal mode, predict the failure risk level according to the abnormal mode, and formulate failure maintenance suggestions according to the failure risk level.

2. The permanent magnet synchronous motor failure monitoring method for the space environment according to claim 1, wherein Construct a simulated space environment test chamber, which includes extreme temperature cycling conditions and high vacuum conditions. Accelerated aging tests are performed on the permanent magnet synchronous motor through the simulated space environment test chamber to obtain multi-dimensional degradation data. The method includes: Execute a stepped temperature cycle according to the extreme temperature cycling conditions of the simulated space environment test chamber, and deploy multi-type sensor arrays at the motor stator winding, permanent magnet surface, and bearing seat of the permanent magnet motor according to the high vacuum conditions of the simulated space environment test chamber to synchronously collect temperature distribution signals, vibration signals, and current harmonic signals; Filter and denoise the temperature distribution signal to obtain motor temperature distribution data, perform empirical mode decomposition on the vibration signal to obtain vibration spectrum data, and transform the current harmonic signal to obtain current harmonic data; Perform time-frequency domain fusion processing on the motor temperature distribution data, the vibration spectrum data, and the current harmonic data to generate the multi-dimensional degradation data.

3. The permanent magnet synchronous motor failure monitoring method for space environment according to claim 2, characterized in that, Based on the multi-dimensional degradation data, perform material property evaluation on the permanent magnet synchronous motor to determine the material property degradation results. The method includes: Based on the motor temperature distribution data, determine the temperature cycle times and perform a demagnetization curve test on the permanent magnet synchronous motor to obtain the permanent magnet demagnetization rate; Based on the vibration spectrum data, perform crystal structure evolution on the bearing lubricating film of the permanent magnet synchronous motor to generate a lubrication failure coefficient; Based on the current harmonic data, calculate the partial discharge inception voltage of the permanent magnet synchronous motor material and draw an insulation performance degradation curve; Couple the permanent magnet demagnetization rate, the lubrication failure coefficient, and the insulation performance degradation curve in multiple physical fields to construct the material property degradation results.

4. The permanent magnet synchronous motor failure monitoring method for space environment according to claim 3, wherein, According to the material property degradation results, conduct failure analysis and construct a space environment failure feature library. The method includes: Based on the permanent magnet demagnetization rate, the lubrication failure coefficient, and the insulation performance degradation curve, perform simulation to determine the prior probabilities of multiple events; Based on the prior probabilities of multiple events, construct a failure tree structure for multiple events, perform weight analysis on the failure tree structure of multiple events, and determine multiple weight coefficients; Based on the failure tree structure of multiple events and the multiple weight coefficients, construct a multi-dimensional failure feature space; Perform clustering analysis according to the multi-dimensional failure feature space to determine multiple clustering centers, where the multiple clustering centers include multiple failure rules, and there is a corresponding relationship between the multiple clustering centers and the multiple failure rules; Extract high-density failure features according to the multiple failure rules to construct the space environment failure feature library.

5. The method for monitoring the failure of a permanent magnet synchronous motor for a space environment according to claim 1, wherein Monitor the motor operation parameter set of the permanent magnet synchronous motor in real time. The method includes: Perform real-time acquisition of the permanent magnet synchronous motor through a high-frequency current probe to obtain three-phase current harmonic component parameters; Perform real-time acquisition of the permanent magnet synchronous motor through a distributed optical fiber temperature sensor to obtain winding temperature gradient parameters; Perform real-time acquisition of the permanent magnet synchronous motor through a vibration sensor to obtain axial vibration acceleration parameters; Perform wavelet packet transform based on the three-phase current harmonic component parameters to extract the characteristic energy values in a fixed frequency band, where the characteristic energy values include current harmonic distortion rate; Perform fluctuation calculation based on the winding temperature gradient parameters to determine the standard deviation of the temperature gradient; Perform feature separation based on the axial vibration acceleration parameters to determine the fretting wear characteristic frequency, where the fretting wear characteristic frequency includes vibration characteristic energy entropy; Integrate the current harmonic distortion rate, the standard deviation of the temperature gradient, and the vibration characteristic energy entropy to construct a composite operation monitoring vector, and add the composite operation monitoring vector to the motor operation parameter set.

6. The permanent magnet synchronous motor failure monitoring method for space environment according to claim 5, wherein Match the motor operation parameter set with the space environment failure feature library to determine the abnormal mode. The method includes: Perform feature dimensionality reduction based on the composite operation monitoring vector to determine a composite dimensionality reduction vector; Use the composite dimensionality reduction vector as an index to traverse the space environment failure feature library for similarity matching to generate multiple matching degrees; When there is a matching degree exceeding the preset warning value among the multiple matching degrees, trigger an online reconfiguration instruction, perform fault analysis through the online reconfiguration instruction, and construct multiple abnormal labels; Backtrack to the motor operation parameter set according to the multiple abnormal labels for identification, and determine the abnormal mode according to the identification result.

7. The method for monitoring the failure of a permanent magnet synchronous motor for a space environment according to claim 1, characterized in that, Predict the failure risk level according to the abnormal mode. The method includes: Extract the bearing wear value and winding insulation deterioration value of the permanent magnet synchronous motor according to the abnormal mode; Perform multivariate degradation failure analysis based on the bearing wear value and the winding insulation deterioration value to determine the joint failure probability; Construct a Markov chain for prediction according to the joint failure probability, construct a risk evolution trend, perform failure impact analysis based on the risk evolution trend, and set the failure risk level.

8. The method for monitoring the failure of a permanent magnet synchronous motor for a space environment according to claim 7, characterized in that, Formulate failure maintenance suggestions according to the failure risk level. The method includes: Set a preset critical value based on the failure risk level, and compare and determine the joint failure probability with the preset critical value; When the joint failure probability exceeds the preset critical value, activate the demagnetization compensation instruction; Perform multi-objective optimization through the demagnetization compensation instruction, construct multi-level failure warning data, perform maintenance control based on the multi-level failure warning data, and formulate the failure maintenance suggestions.

9. A permanent magnet synchronous motor failure monitoring system for the space environment, characterized in that, The system is used to implement the method for monitoring the failure of a permanent magnet synchronous motor for a space environment according to any one of claims 1-8. The system includes: A multi-dimensional degradation data acquisition module is used to construct a simulated space environment experimental chamber. The simulated space environment experimental chamber includes extreme temperature cycling conditions and high vacuum conditions. An accelerated aging experiment is carried out on the permanent magnet synchronous motor through the simulated space environment experimental chamber to obtain multi-dimensional degradation data; A feature library construction module is used to evaluate the material properties of the permanent magnet synchronous motor based on the multi-dimensional degradation data, determine the material property degradation results, conduct failure analysis according to the material property degradation results, and construct a space environment failure feature library; A failure maintenance recommendation formulation module is used to continuously monitor the set of motor operation parameters of the permanent magnet synchronous motor, match the set of motor operation parameters with the space environment failure feature library, determine the abnormal mode, predict the failure risk level according to the abnormal mode, and formulate failure maintenance recommendations according to the failure risk level.

10. An electronic device, characterized in that, The electronic device includes: A processor; A memory for storing executable instructions of the processor; Wherein, the processor is used to execute the method for monitoring the failure of the permanent magnet synchronous motor for space environment according to any one of claims 1 to 8.

Images