Antenna servo motor fault diagnosis method and device for satellite communication, and medium

Abstract

The invention discloses an antenna servo motor fault diagnosis method and device for satellite communication and a medium, and relates to the technical field of satellite communication. The method comprises the following steps: acquiring vibration signals, current signals and AGC signals of a servo motor under various working conditions; the current signals are screened based on the AGC signals, and non-fault current signals caused by satellite interference factors are removed; performing feature extraction on the vibration signals and the screened current signals to obtain training features; training a neural network model by using the training features to obtain a fault diagnosis model of the antenna servo motor; and performing automatic fault diagnosis on the antenna servo motor through the trained fault diagnosis model. According to the method, the accuracy, robustness and reliability of fault detection are improved through multi-sensor information fusion, and misjudgment and missed judgment are reduced.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication technology, and in particular to a method, equipment and medium for diagnosing faults of a satellite communication antenna servo motor. Background Art

[0002] The antenna servo motor is the core part of the ground measurement and control system that drives the antenna to track the satellite. Its stable operating state is the key to ensuring that the ground system can receive satellite signals with high quality.

[0003] In a satellite communication scenario, when a satellite sends a signal to a ground antenna, the communication equipment on the satellite converts the data to be transmitted (satellite status data, various detected space information, etc.) into electromagnetic waves suitable for propagation in free space and sends them out. These signals propagate toward the earth in the form of radio waves, but due to the attenuation of the signal during propagation, the signal strength received on the ground may be very weak. In order to cope with this signal attenuation, the automatic gain control (AGC) system is designed to automatically adjust the gain of the received signal and modulate the amplitude of the satellite output signal at an appropriate level. The antenna servo motor in the ground measurement and control system is responsible for driving the antenna to align with the satellite, ensuring that the data sent by the satellite is received and avoiding the loss of telemetry data. However, any failure of the servo motor may cause the antenna to fail to point correctly, thereby affecting the reception of telemetry data and the tracking of the satellite.

[0004] With the increasing number of space launch and measurement and control missions, higher requirements are placed on the stability of ground equipment operation. Antenna servo motor fault diagnosis technology can prevent potential failures and ensure the continuity and reliability of satellite communications, which is crucial to maintaining the high-performance operation of satellite communication systems.

[0005] Traditional fault diagnosis methods mainly rely on expert experience, manual regular inspection and maintenance. These methods are subjective and limited and cannot meet the needs of fast and accurate diagnosis. With the development of hardware technologies such as GPU and the popularization of Internet and sensor technology, there are a large number of methods for diagnosing motor faults based on machine learning and deep learning. For example, some researchers use neural networks to diagnose faults in radar antenna servo systems, while others establish equivalent motor models for antenna servo systems and combine them with deep learning methods for fault diagnosis. Machine learning-based fault diagnosis methods usually extract fault features from collected data by establishing and training machine learning models, select sensitive features and train models for fault diagnosis; commonly used methods for deep learning-based fault diagnosis include autoencoders, deep belief networks, convolutional neural networks, and recurrent neural networks.

[0006] However, the existing fault diagnosis based on vibration signal or current signal analysis still has limitations. Since vibration signals are easy to collect and contain a lot of data status information, most of the existing motor fault diagnosis methods are studied around vibration signals. However, vibration signals are prone to contain noise and coupled vibration information of other components, which will affect the complexity of the system and the accuracy of diagnosis. Current signals are favored by researchers because of their advantages such as easy collection, high signal sensitivity and strong stability. However, the fault information contained in traditional single sensors is limited and often cannot accurately reflect the intrinsic characteristics of the current state of the antenna servo motor. In particular, when the sensor fails or cannot reflect the fault characteristics, misjudgment may occur.

[0007] In summary, in the diagnosis of the antenna servo motor of the ground station antenna tracking satellite system, the fault information reflected by a single signal is one-sided, and the signal acquisition process is easily affected by factors such as sensor failure and environmental noise, resulting in problems such as misjudgment and missed judgment. Summary of the invention

[0008] The present invention provides a satellite communication antenna servo motor fault diagnosis method, equipment and medium to solve the problem that a single signal reflects one-sided fault information and is easily misjudged or missed due to the influence of a signal acquisition sensor.

[0009] The present invention is achieved through the following technical solutions:

[0010] A first aspect of the present invention provides a method for diagnosing a fault of a satellite communication antenna servo motor, comprising:

[0011] Acquire vibration signals, current signals and AGC signals of the antenna servo motor under various working conditions, wherein the various working conditions include operating conditions under normal conditions and various fault conditions; the antenna servo motor is a servo motor that drives the antenna to track the satellite, and the AGC signal is an automatic gain control signal received by the antenna during the satellite tracking process;

[0012] The current signal is screened based on the AGC signal to remove non-fault current signals caused by satellite interference factors; the vibration signal and the screened current signal are subjected to feature extraction to obtain training features;

[0013] Using the training features to train a neural network model to obtain a fault diagnosis model for an antenna servo motor;

[0014] Fault diagnosis of the antenna servo motor is performed based on the fault diagnosis model.

[0015] The present invention starts from the perspective of multi-source signal fusion, fuses the vibration signal, current signal and AGC signal from the satellite signal receiving process of the antenna servo motor, and applies them to the training of the fault diagnosis model, wherein the AGC signal is used to screen the fault current signal. The AGC signal is a numerical sequence that is automatically calculated and generated by the satellite receiving system based on the satellite signal strength and used to dynamically adjust the gain. In the tracking process of the satellite passing, the measurement and control antenna receives the value. When the AGC value fluctuates greatly, the antenna tracking is unstable, which may cause the motor current signal to fluctuate, thereby affecting the fault diagnosis of the motor. Therefore, by screening the AGC signal, it is possible to avoid confusing the non-fault current information caused by interference factors such as unstable satellite signals with sample information, thereby avoiding the interference model's learning of fault features. The vibration signals and current signals collected by different source sensors are used together to characterize the health status of the antenna servo motor, and feature samples are constructed as key information for fault diagnosis, and the antenna servo motor fault diagnosis model is trained to avoid the problem of one-sided characterization of a single data source. Thereby improving the accuracy, robustness and reliability of fault detection, reducing misjudgment and missed judgment.

[0016] In one implementation, the operating conditions under the multiple fault states include at least the operating conditions of the antenna servo motor under the conditions of bearing damage, stator winding inter-turn short circuit and air gap eccentricity.

[0017] In one embodiment, the screening of the current signal based on the AGC signal to remove non-fault current signals caused by satellite interference factors includes: calculating the variance of the AGC signal and removing the current signal when the variance of the AGC signal is greater than a set threshold.

[0018] In one embodiment, extracting features from the vibration signal and the filtered current signal to obtain training features includes:

[0019] De-noising the vibration signal and the filtered current signal respectively;

[0020] Extract energy features from the denoised vibration signal, and extract multiple power frequency amplitude features from the denoised current signal;

[0021] The energy feature of the vibration signal and the multiple power frequency amplitude feature of the current signal are spliced ​​to obtain the training feature.

[0022] In one embodiment, the extracting energy features of the denoised vibration signal includes: performing three-layer wavelet decomposition on the denoised vibration signal, reconstructing each sub-frequency band obtained by the decomposition, calculating the energy value of each reconstructed signal, and normalizing the energy value to obtain the energy feature of the vibration signal.

[0023] In one embodiment, the extracting of multiple power frequency amplitude features of the denoised current signal includes: performing wavelet packet transform on the denoised current signal to obtain the amplitude of each multiple power frequency, and combining the amplitude of each multiple power frequency into multiple power frequency amplitude features; wherein each multiple power frequency includes at least 1 / 2 times the power frequency, 1 times the power frequency, 2 / 3 times the power frequency, 2 times the power frequency, 3 times the power frequency and 4 times the power frequency.

[0024] In one embodiment, the neural network model includes a Transformer encoder module, a pooling layer, a fully connected layer and an output layer; the Transformer encoder module is formed by six stacked Transformer Encoders, each Transformer Encoder includes a set of multi-head attention networks and a feedforward network, the output of the multi-head attention is input to the feedforward network through residual and normalization, and the output of the feedforward network is input to the pooling layer through residual and normalization.

[0025] In one implementation, the performing fault diagnosis on the antenna servo motor based on the fault diagnosis model includes:

[0026] Acquire a vibration signal and a current signal of an antenna servo motor to be diagnosed, wherein the antenna servo motor is a servo motor that drives the antenna to track a satellite;

[0027] Extracting features from the vibration signal and the current signal respectively, and splicing the extracted features;

[0028] The spliced ​​features are input into the fault diagnosis model to obtain the diagnosis result.

[0029] The second aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method for diagnosing antenna servo motor faults in satellite communications as described in any one of the first aspects of the present invention.

[0030] According to a third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for diagnosing antenna servo motor faults in satellite communications as described in any one of the first aspects of the present invention is implemented.

[0031] Compared with the prior art, the present invention has the following advantages and beneficial effects: from the perspective of multi-source signal fusion, the vibration signal, current signal of the antenna servo motor, and the AGC signal from the satellite signal reception process are fused for fault identification model training. The AGC signal is used to screen the fault current information, the vibration signal and current signal collected by the sensor are used to characterize the health status of the antenna servo motor, and provide key information for fault diagnosis. The vibration signal and current signal are used to construct a feature sample data set, and the antenna servo motor fault diagnosis model is trained based on this data set. Finally, the antenna servo motor is fault diagnosed through the trained model. Multi-sensor information fusion is used to improve the accuracy, robustness and reliability of fault detection, improve the accuracy of antenna servo motor fault diagnosis, and reduce misjudgment and missed judgment in antenna servo motor fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0033] Figure 1 It is a flow chart of a method for diagnosing a fault of a satellite communication antenna servo motor according to an embodiment of the present invention;

[0034] Figure 2 is a flow chart of a multi-source data feature extraction method according to an embodiment of the present invention;

[0035] Figure 3 It is a structural schematic diagram of a satellite antenna servo motor fault diagnosis model according to an embodiment of the present invention;

[0036] Figure 4 It is a flow chart of another method for diagnosing faults of a satellite antenna servo motor according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0038] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to or inherent to other steps or units of the device.

[0039] The terms used in the various embodiments of the application are only used to describe the purpose of specific embodiments and are not intended to limit the various embodiments of the application. As used herein, the singular form is intended to also include the plural form, unless the context clearly indicates otherwise. Unless otherwise limited, all terms used here (including technical terms and scientific terms) have the same meaning as the meanings commonly understood by ordinary technicians in the field of the various embodiments of the application. The terms (such as the terms defined in the dictionary generally used) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or an overly formal meaning, unless clearly defined in the various embodiments of the application.

[0040] The present invention proposes a method, device and medium for diagnosing antenna servo motor faults in satellite communications. The diagnostic method collects vibration signals and current signals of the antenna servo motor and satellite AGC signals received by the measurement and control antenna by arranging different types of sensors. Multi-sensor information fusion is used to improve the accuracy, robustness and reliability of fault detection, improve the accuracy of antenna servo motor fault diagnosis, and reduce misjudgment and missed judgment.

[0041] like Figure 1 Shown is a flow chart of a method for diagnosing faults of a satellite communication antenna servo motor according to an embodiment of the present invention, which includes the following steps.

[0042] S1: Obtain the vibration signal, current signal and AGC signal of the antenna servo motor under various working conditions.

[0043] The vibration signal and current signal of the antenna servo motor, as well as the satellite AGC (dynamic gain adjustment) signal received by the measurement and control antenna are collected through vibration sensors and current sensors. The AGC signal is a numerical sequence based on the satellite signal strength, which is automatically calculated and generated by the satellite receiving system and used to dynamically adjust the gain.

[0044] In the present invention, the antenna servo motor is a servo motor that drives the antenna to track the satellite, and the AGC signal is an automatic gain control signal received by the antenna during the process of tracking the satellite.

[0045] The vibration signals, current signals and AGC signals collected under various working conditions are used as training samples to train the fault diagnosis model. The training samples include the collected data of the antenna servo motor in the fault state and the normal state, and the corresponding working conditions are used as sample labels.

[0046] In one embodiment, the multiple fault conditions include at least common faults such as bearing damage, stator winding inter-turn short circuit, air gap eccentricity, etc. In practical applications, more fault conditions can also be added. By simulating the fault, the vibration signal and current signal of the antenna servo motor in the fault state and the AGC signal received by the antenna in the operating state are collected to form fault samples. These samples can come from the same ground station or different ground stations and different antenna servo motors.

[0047] Furthermore, according to the small sample principle, the signals collected under various working conditions are divided into training sets and test sets, and the training sets and test sets are randomly scattered. The training set is used to train the fault recognition model, and the test set is used to test the model, adjust the model parameters, and improve the model performance.

[0048] S2: Filter the current signal based on the AGC signal to remove the non-fault current signal caused by satellite interference factors.

[0049] The current of the motor during operation is affected by many factors, including motor parameters, AGC values, etc. There is a correlation between the AGC signal and the motor current signal. The AGC value reflects the satellite signal-related situation, and its stability is related to the change of the motor current. When the AGC value fluctuates greatly, the antenna tracking is unstable, which will interfere with the motor current and affect the fault diagnosis results. The satellite AGC signal is used to avoid confusing non-fault current information with fault sample information due to interference factors such as unstable satellite signals, and eliminate it to avoid interfering with the model's learning of fault features.

[0050] By calculating the variance of the AGC signal, the current signal when the AGC signal variance is greater than the set threshold k is removed. Specifically, during the process of the antenna tracking the satellite, the start time, end time and satellite AGC value received by the measurement and control antenna are recorded, and then the signal that satisfies the AGC signal variance greater than the threshold k during this time period is removed. This threshold k can be based on experience or the optimal value obtained by in-depth analysis and model training of a large amount of historical data. The training process involves data under various operating conditions of the antenna servo motor, including normal state, different fault states, etc.

[0051] The screened current signal is the steady-state current signal of the antenna servo motor, which can more accurately reflect the actual operating status of the motor. The fault diagnosis model is trained with multi-source data consisting of the screened steady-state current signal and the collected vibration signal under different working conditions, avoiding the mixing of non-fault current data caused by interference factors such as unstable satellite signals into the current sample data used for analysis and diagnosis, thereby improving the accuracy of fault diagnosis.

[0052] S3: Extract features from the vibration signal and the filtered current signal to obtain training features.

[0053] In one implementation, the energy feature of the vibration signal and the multiple power frequency amplitude feature of the current signal are used to fuse the two features and use them as input features for training the fault diagnosis model. Figure 2 The flowchart of the multi-source data feature extraction method shown in the figure first performs denoising on the vibration signal and the filtered current signal respectively; then extracts the energy feature of the denoised vibration signal, and extracts the multiple power frequency amplitude feature of the denoised current signal; finally, concatenates the energy feature of the vibration signal and the multiple power frequency amplitude feature of the current signal to obtain the training feature.

[0054] In one implementation, the denoising step uses wavelet soft thresholding to perform denoising.

[0055] In one embodiment, extracting energy features from a denoised vibration signal specifically includes: performing three-layer wavelet decomposition on the denoised vibration signal, reconstructing each sub-frequency band obtained by the decomposition, calculating the energy value of each reconstructed signal, and normalizing the energy value to obtain the energy feature of the vibration signal.

[0056] The denoised vibration signal is decomposed into three layers of wavelet packets, and the formula is:

[0057]

[0058] Among them, h k is a high-pass filter, g k is a low-pass filter, x n (2t-k) is the input signal of the filter, x 2n (t) is the high-pass filter component, x 2n+1 (t) is the low-pass filter component, t represents the sampling time, k represents the index of the filter coefficient, k∈Z, Z represents the index value range, such as 0, 1, 2... The signal passes through h k , g k Two filters perform two samples at the same time, decomposing the original signal into high-frequency and low-frequency parts.

[0059] Each time the signal is decomposed, the data of each frequency band is halved, and the bandwidth of the wavelet packet decomposition Δf, the number of decomposition layers j, and the sampling frequency f s Meet the following requirements:

[0060] Δf=fs s / (2*2 j )

[0061] After three layers of wavelet packet decomposition, 8 sub-band signals are finally obtained. The sub-band signals obtained by the third layer of wavelet packet decomposition are reconstructed and the energy value of each sub-band signal is calculated.

[0062] The steps to calculate the band energy value of three-layer wavelet packet decomposition are as follows:

[0063] Let X be the vibration signal after denoising, (i,j) be the layer node after wavelet packet decomposition, i.e., the jth node of the i-th layer, i = (0,1,2,…,N), j = (0,1,2,…,2 N -1), N is the number of decomposed layers, and in this embodiment, N=3.

[0064] Reconstruct the wavelet of each frequency band, using X1,X2,…,X n Represents the reconstructed signal of each frequency band, and the energy value of the reconstructed signal of each frequency band is calculated as:

[0065]

[0066] Among them, x ik Represents the reconstructed signal X i The amplitude of the kth discrete point, e i Represents the reconstructed signal X i The energy value of , m represents the total number of samples.

[0067] After the calculated energy values ​​are normalized, the characteristic matrix E1 of the vibration signal is formed, where the normalization formula is:

[0068]

[0069] The energy feature matrix of the final vibration signal is F1 = [E1, E2, E3, E4, E5, E6, E7, E8].

[0070] In one embodiment, extracting multiple power frequency amplitude features from the denoised current signal specifically includes: performing wavelet packet transform on the denoised current signal to obtain the amplitudes of each multiple power frequency, and combining the amplitudes of each multiple power frequency into multiple power frequency amplitude features.

[0071] The spectrum analysis of the current signal through wavelet packet transform is similar to the analysis steps of the vibration signal. After obtaining the amplitude of each frequency component, 1 / 2 times the power frequency amplitude, 1 times the power frequency amplitude, 2 / 3 times the power frequency amplitude, 2 times the power frequency amplitude, 3 times the power frequency amplitude, and 4 times the power frequency amplitude are selected to form the current multiple power frequency amplitude matrix F2 = [Z1, Z2, Z3, Z4, Z5, Z6].

[0072] Finally, the energy feature F1 of the vibration signal and the multiple power frequency amplitude feature F2 of the current signal are spliced ​​to form a fusion feature matrix F = [F1, F2] = [E1, E2, E3, E4, E5, E6, E7, E8, Z1, Z2, Z3, Z4, Z5, Z6].

[0073] S4: Use the training features to train the neural network model to obtain a fault diagnosis model for the antenna servo motor.

[0074] The fault diagnosis model of the antenna servo motor in the present invention is built based on a neural network model. The fusion feature matrix F obtained in step S3 is input into the fault diagnosis model to obtain a prediction result, and the model parameters are trained based on the gap between the prediction result and the sample label (working condition: normal / fault type).

[0075] In one embodiment, the neural network model adopts a Transformer-based deep neural network, including a Transformer encoder module, a pooling layer, a fully connected layer and an output layer.

[0076] See also Figure 3 The schematic diagram of the structure of the antenna servo motor fault diagnosis model is shown in Figure 1. The Transformer encoder module is formed by six stacked Transformer Encoders, each of which includes a set of multi-head attention networks (MSA) and feedforward networks (FFN). Residual connections are used between the MSA and FFN layers, and normalization (LN) is used after the residual connection. The output of the feedforward network is connected to the residual and normalization. The above structures together constitute TransformerEncode. The output of the multi-head attention is input to the feedforward network through the residual and normalization, and the output of the feedforward network is input to the pooling layer through the residual and normalization.

[0077] In the multi-head self-attention mechanism, for the input matrix Using different initialization matrices to linearly calculate the key, query, and value matrices, in each head (K, Q, V), the attention mechanism is executed in parallel, and finally produces d v dimensional output value. The weighted score matrix of the attention mechanism is calculated as follows:

[0078] Given a query matrix A key matrix and a matrix of values The self-attention is calculated by the following formula:

[0079]

[0080] Among them, d is the dimension of the query matrix and the key matrix, and the query matrix Q, key matrix K, and value matrix V are learnable parameters. Multi-head attention connects multiple of the above formulas together. The output of the multi-head attention is input to the feedforward network through residual and normalization, and the output of the feedforward network is input to the pooling layer through residual and normalization. The calculation process of MSA and FFN is as follows:

[0081] F o =F i +MSA(LN(F i ))

[0082] F i+1 =F o +FFN(LN(F o ))

[0083] Among them, F i and F o They represent the input and output of the i-th layer respectively, i=6, and the Transformer encoder module is formed by stacking six encoders.

[0084] The fused feature matrix F is input into the Transformer encoder module to extract key fault features. The encoder output is input into the pooling layer, and global average pooling is used to reduce the feature dimension and compress the feature information. The pooled features are input into the fully connected layer for further processing. The fully connected layer maps the pooled features to the category space through linear transformation and activation function. The output layer uses the Softmax classifier to classify the extracted fault features and obtain the classification results of the fault data.

[0085] In this embodiment, the cross entropy loss function is selected for model training to measure the gap between the model prediction result and the actual label. The cross entropy classification loss L class The mathematical expression is:

[0086]

[0087] Among them, i represents the category index, y i and Represent the actual probability distribution and predicted probability distribution of the model respectively.

[0088] After defining the loss function, the Adam optimizer is used to update the model parameters through the back-propagation algorithm to minimize the loss function.

[0089] Furthermore, the model trained by the above method is further adjusted using a test set to obtain a final fault diagnosis model.

[0090] like Figure 4 Shown is a flow chart of a fault diagnosis method for an antenna servo motor, which includes the following steps.

[0091] S11, obtaining a vibration signal and a current signal of the antenna servo motor to be diagnosed.

[0092] S12, extracting features from the vibration signal and the current signal respectively, and concatenating the extracted features.

[0093] S13, inputting the spliced ​​features into the fault diagnosis model of the antenna servo motor to obtain a diagnosis result.

[0094] The current signal of the antenna servo motor to be diagnosed collected in the above step S11 is a steady-state current signal of the antenna servo motor, that is, collected without being disturbed by unstable satellite signals, thereby improving the accuracy of the diagnosis result. The steps of extracting features from the vibration signal and the current signal in step S12 are referred to the above embodiment.

[0095] An embodiment of the present invention further provides a fault diagnosis model training device for an antenna servo motor, comprising:

[0096] An acquisition module is used to acquire vibration signals, current signals and AGC signals of the antenna servo motor under various working conditions;

[0097] A multi-source data processing module is used to screen the current signal based on the AGC signal and remove the non-fault current signal caused by satellite interference factors;

[0098] A feature extraction module is used to extract features from the vibration signal and the filtered current signal to obtain training features;

[0099] The training module is used to train the neural network model using the training features to obtain a fault diagnosis model for the antenna servo motor.

[0100] The antenna servo motor is a servo motor that drives the antenna to track the satellite, and the AGC signal is a satellite automatic gain control signal received by the antenna during the satellite tracking process. The execution method of the multi-source data processing module, the feature extraction module and the training module can be seen from the above embodiments.

[0101] An embodiment of the present invention further provides an electronic device, which includes a processor and a memory, and the number of processors can be one or more. The memory, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules. The processor executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory, so as to implement the antenna servo motor fault diagnosis method for satellite communication of any of the above embodiments of the present invention.

[0102] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0103] In some examples, the memory may further include a memory remotely located relative to the processor, and these remote memories may be connected to the electronic device via a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0104] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for diagnosing a fault of a satellite communication antenna servo motor according to any of the above embodiments of the present invention is implemented. The storage medium may be a ROM / RAM, a magnetic disk, an optical disk, etc.

[0105] An embodiment of the present invention further provides a computer program product. When the computer program product is run on a computer, the computer is enabled to execute the method for diagnosing faults of a satellite communication antenna servo motor in any of the above embodiments of the present invention.

[0106] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for diagnosing faults of a satellite communication antenna servo motor, characterized in that: include: Acquire vibration signals, current signals and AGC signals of the antenna servo motor under various working conditions, wherein the various working conditions include operating conditions under normal conditions and various fault conditions; The antenna servo motor is a servo motor that drives the antenna to track the satellite, and the AGC signal is an automatic gain control signal received by the antenna during the process of tracking the satellite; The current signal is screened based on the AGC signal to remove the non-fault current signal caused by satellite interference factors; Extracting features from the vibration signal and the filtered current signal to obtain training features; Using the training features to train a neural network model to obtain a fault diagnosis model for an antenna servo motor; Fault diagnosis of the antenna servo motor is performed based on the fault diagnosis model.

2. The method for diagnosing a fault of a satellite communication antenna servo motor according to claim 1, characterized in that: The operating conditions under the multiple fault states include at least the operating conditions of the antenna servo motor under the conditions of bearing damage, stator winding inter-turn short circuit and air gap eccentricity.

3. The method for diagnosing faults of a satellite communication antenna servo motor according to claim 1, characterized in that: The screening of the current signal based on the AGC signal to remove the non-fault current signal caused by the satellite interference factor includes: calculating the variance of the AGC signal and removing the current signal when the variance of the AGC signal is greater than a set threshold.

4. The method for diagnosing faults of a satellite communication antenna servo motor according to claim 1, characterized in that: The step of extracting features from the vibration signal and the filtered current signal to obtain training features includes: De-noising the vibration signal and the filtered current signal respectively; Extract energy features from the denoised vibration signal, and extract multiple power frequency amplitude features from the denoised current signal; The energy feature of the vibration signal and the multiple power frequency amplitude feature of the current signal are spliced ​​to obtain the training feature.

5. The method for diagnosing faults of a satellite communication antenna servo motor according to claim 4, characterized in that: The method of extracting energy features from the denoised vibration signal comprises: performing three-layer wavelet decomposition on the denoised vibration signal, reconstructing each sub-frequency band obtained by the decomposition, calculating the energy value of each reconstructed signal, and normalizing the energy value to obtain the energy features of the vibration signal.

6. The method for diagnosing faults of a satellite communication antenna servo motor according to claim 4, characterized in that: The method of extracting multiple power frequency amplitude features from the denoised current signal includes: performing wavelet packet transform on the denoised current signal to obtain the amplitudes of each multiple power frequency, and combining the amplitudes of each multiple power frequency into multiple power frequency amplitude features; wherein each multiple power frequency includes at least 1 / 2 times the power frequency, 1 times the power frequency, 2 / 3 times the power frequency, 2 times the power frequency, 3 times the power frequency and 4 times the power frequency.

7. The method for diagnosing faults of a satellite communication antenna servo motor according to claim 1, characterized in that: The neural network model includes a Transformer encoder module, a pooling layer, a fully connected layer and an output layer; the Transformer encoder module is formed by six stacked Transformer Encoders, each Transformer Encoder includes a set of multi-head attention networks and a feedforward network, the output of the multi-head attention is input to the feedforward network through residual and normalization, and the output of the feedforward network is input to the pooling layer through residual and normalization.

8. The method for diagnosing faults of a satellite communication antenna servo motor according to claim 1, characterized in that: The method of performing fault diagnosis on the antenna servo motor based on the fault diagnosis model includes: Acquire a vibration signal and a current signal of an antenna servo motor to be diagnosed, wherein the antenna servo motor is a servo motor that drives the antenna to track a satellite; Extracting features from the vibration signal and the current signal respectively, and splicing the extracted features; The spliced ​​features are input into the fault diagnosis model to obtain the diagnosis result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the satellite communication antenna servo motor fault diagnosis method described in any one of claims 1-8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the satellite communication antenna servo motor fault diagnosis method described in any one of claims 1-8 is implemented.

Images