Radar servo system vibration characteristic online diagnosis and evaluation method

Abstract

The invention belongs to the technical field of radar system comprehensive guarantee, and discloses a radar servo system vibration characteristic on-line diagnosis and evaluation method, which comprises the following steps of: extracting a frequency domain characteristic value and a time domain characteristic value according to radar servo vibration sensor monitoring data received in real time, performing fault analysis and performance evaluation in two dimensions of frequency and time domain, and determining the vibration characteristic of a radar servo system. And the health state of the current radar servo system is obtained by combining the frequency domain and time domain characteristic value intervals generated during normal vibration, so that the monitoring and diagnosis blind areas of the servo system are weakened, the requirement on the high technical level of a user is reduced, and the operation reliability of the radar system is improved.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of comprehensive support for radar systems, and particularly relates to a method for on-line diagnosis and evaluation of the vibration characteristics of a radar servo system. Background Art

[0002] As one of the key devices of a radar, the radar servo system plays a crucial role in the radar's detection mission. It not only controls the pointing and movement of the radar antenna, but also directly affects the detection accuracy, response speed, and stability of the radar. With the rapid development of modern radar technology, the performance requirements for the servo system of the radar system are increasing day by day. Especially in complex environments and high-speed movement conditions, the vibration characteristics of the servo system are directly related to the overall performance of the radar. Therefore, there is an urgent need for a method that can quickly and accurately diagnose and evaluate the vibration characteristics of the radar servo system on-line to ensure the efficient operation and reliability of the radar system.

[0003] Currently, the diagnosis and evaluation of the radar servo system mainly rely on manual experience judgment. Although this method can solve some obvious fault problems to a certain extent, its limitations are also very obvious. First, manual judgment is too dependent on the experience level of the operator. When the fault manifestation is not obvious or the system performance gradually deteriorates, it is very difficult to make an accurate diagnosis based on manual experience alone. Second, manual judgment lacks the support of quantitative characteristic values and cannot objectively and accurately evaluate the current performance of the servo system, resulting in users being difficult to accurately grasp the health status of the radar servo system. In addition, with the increase in the complexity of the radar system, the fault modes of the servo system have become more diverse and hidden, and simply relying on manual judgment can no longer meet the maintenance requirements of modern radar systems.

[0004] At the same time, a large amount of vibration sensor monitoring data is generated in real time during the operation of the radar servo system. These data contain rich system state information, but due to the large amount of data and complex information, how to efficiently use these data for rapid analysis, feature extraction, and diagnostic evaluation has become the main problem faced by the current diagnosis and evaluation of the vibration characteristics of the radar servo system. The traditional off-line analysis method not only takes a long time, but also cannot meet the real-time requirement, and it is difficult to detect problems in time during the operation of the system and take corresponding measures.

[0005] Therefore, developing an online diagnosis and evaluation method based on data-driven has become the key to solving this difficult problem. By introducing advanced signal processing techniques, machine learning algorithms, and big data analysis methods, key features reflecting the health status of the servo system can be extracted from a large amount of vibration monitoring data, and corresponding diagnostic models can be established. This method can not only achieve real-time monitoring and rapid diagnosis of the vibration characteristics of the servo system, but also provide quantitative performance evaluation indicators to help users accurately grasp the health status of the system, take maintenance measures in a timely manner, and thus effectively improve the reliability and service life of the radar system. Summary of the Invention

[0006] The purpose of the present invention is to provide an online diagnosis and evaluation method for the vibration characteristics of a radar servo system, aiming to extract frequency domain eigenvalue and time domain eigenvalue according to the monitoring data of the radar servo vibration sensor received in real time, conduct fault analysis and performance evaluation in two dimensions of frequency and time domain, and combine the frequency domain and time domain eigenvalue intervals generated during normal vibration to obtain the health status of the current radar servo system, weakening the monitoring and diagnosis blind spots of the servo system, reducing the requirements for high technical levels of users, and improving the reliability of the radar system operation.

[0007] To achieve the above purpose, the present invention provides an online diagnosis and evaluation method for the vibration characteristics of a radar servo system, including the following steps: Step 1: Receive the vibration monitoring data of the radar servo system. The vibration monitoring data includes vibration frequency domain data and vibration time domain data. First, preprocess the vibration frequency domain data and vibration time domain data, and switch to the vibration frequency domain data or vibration time domain data processing channel according to the data type. Step 2: If entering the vibration frequency domain data processing channel, first determine whether the current system speed is greater than or equal to the set threshold A; If the speed is less than A, discard the packet of vibration frequency domain data and continue to monitor the data; If the speed is greater than or equal to A, add the vibration frequency domain data to the buffer area, and determine whether the current buffer area data volume is greater than or equal to the set threshold B; Step 3: If the buffer area data volume is less than B, continue to collect frequency domain data; If the buffer area data volume is greater than or equal to B, obtain the maximum amplitude and the frequency point information at the maximum amplitude from the parsed frequency domain waveform, and obtain the normal frequency point sets in the X / Y / Z three dimensions according to the current speed and position; Step 4: Compare and analyze with the normal frequency point set. If an abnormality occurs, add it to the abnormal set, continue to process the next packet of frequency domain data until all the data in the buffer area is processed, and conduct servo comprehensive diagnosis and evaluation; Step 5: If entering the vibration time domain data processing channel, first determine whether the current system speed is greater than or equal to the set threshold A; If the rotational speed is less than A, discard the vibration time-domain data of this packet and continue to monitor the data; If the rotational speed is greater than or equal to A, add the vibration time-domain data to the buffer, and determine whether the current data volume in the buffer is greater than or equal to the set threshold C; Step 6: If the data volume in the buffer is less than C, continue to collect time-domain data; If the data volume in the buffer is greater than or equal to C, calculate the kurtosis values and root mean square values in the three dimensions of X / Y / Z, and obtain the normal kurtosis value and root mean square value threshold information in the three dimensions of X / Y / Z according to the current rotational speed and position; Step 7: Compare and analyze with the normal kurtosis value and root mean square value thresholds. If an abnormality occurs, add it to the abnormal set, continue to process the time-domain data of the next packet until all the data in the buffer is processed, and perform servo comprehensive diagnosis and evaluation.

[0008] Further, the range of the threshold A is 3 - 6 rpm, the range of the threshold B is 120 - 150, and the range of the threshold C is 3000 - 3200.

[0009] Further, the threshold A set in Steps 2 and 5 is 3 rpm, the threshold B set in Steps 2 and 3 is 120, and the threshold C set in Steps 5 and 6 is 3200.

[0010] Further, it further includes Step 8: Enter the servo comprehensive diagnosis and evaluation process: Calculate the frequency point error. If it is an occasional abnormality, eliminate it. If it is not an occasional abnormality, report the vibration frequency-domain fault; Calculate the kurtosis value and root mean square value error. If an abnormality occurs, report the vibration time-domain fault.

[0011] Further, it further includes Step 9: Based on the comprehensive analysis of the servo vibration frequency-domain and time-domain faults and characteristic values, judge the health level of the system.

[0012] Further, in Step 1, preprocess the vibration time-domain data and the vibration time-domain data, and use the empirical mode decomposition EMD noise reduction algorithm based on the intrinsic mode function IMF for data preprocessing.

[0013] Further, the preprocessing of the vibration time-domain data and the vibration time-domain data in Step 1 includes the following steps: S1: First, decompose the original vibration monitoring data into the sum of multiple IMF components and a residual component. The formula is , where L is the number of IMF components, is the i-th IMF component, R(t) is the residual component, and is a non-zero mean slow-varying function composed of a few extreme values; S2: Then, perform Gaussian white noise test for each IMF component. The formula is , where is the autocorrelation sequence calculation operator, is the corresponding autocorrelation function, and the first item of the result is taken and the absolute value is used as the difference degree from white noise; S3: If , it is considered that the difference is small, and this IMF component is noise and should be removed; S4: If , it is considered that the difference is large, and the part of the intrinsic mode components with a large difference from Gaussian white noise is selected for reconstruction or summation to obtain the denoised data.

[0014] Further, the IMF component should meet the following conditions: For the entire vibration monitoring data set, the number of zero-crossing points and extreme points must be equal or differ by at most one; At any time, the average value of the envelope formed by local maxima and minima should be zero.

[0015] Further, the calculation formula for the kurtosis value in step 6 is , where ; The calculation formula for the root mean square value is ; where is the vibration amplitude value, is the number of sampling points.

[0016] Further, according to the comprehensive analysis of the servo vibration frequency domain and time domain faults and characteristic values, it is divided into the following 4 cases: T1: If the abnormal sets are all empty, it means that the servo state is normal at this time, there is no fault in the servo system, and the health level of the radar servo system is evaluated as normal; T2: If the abnormal set calculated in the frequency domain is not empty, the evaluation level needs to be determined according to the abnormal situation, and the error value between the abnormal frequency point and the normal frequency point is calculated. The calculation formula is , where is the abnormal frequency, is the normal frequency; When the error value is greater than the frequency error tolerance value and the occurrence times are greater than 50, it is determined as a vibration frequency domain fault, and the health level of the radar servo system is evaluated as inefficient; When the occurrence times are less than 50, it is determined as an occasional anomaly; T3: If the abnormal set calculated in the time domain is not empty, then according to the kurtosis value and root mean square value abnormal sets calculated in the time domain, the error value of the characteristic values in the abnormal set is calculated. The formula is , , where, when is the case, , when , ; when , , when , ; When the error value is greater than the amplitude error tolerance value, it is determined as a vibration time-domain fault, and the health level of the radar servo system is evaluated as inefficient; T4: If faults occur in both the vibration frequency domain and time domain, then the health level of the radar servo system is evaluated as failed.

[0017] Beneficial effects: The present invention provides an online diagnosis and evaluation method for the vibration characteristics of a radar servo system. According to the monitoring data of the radar servo vibration sensor received in real time, through advanced signal processing techniques, frequency-domain characteristic values and time-domain characteristic values are extracted, and a comprehensive fault analysis and performance evaluation of the radar servo system are carried out from two dimensions of frequency and time domain. By comparing and analyzing the extracted characteristic values with the preset frequency-domain and time-domain characteristic value intervals under normal vibration conditions, the current health state of the radar servo system can be accurately judged, and potential fault modes or performance degradation trends can be identified.

[0018] This method effectively weakens the blind areas existing in traditional monitoring and diagnosis. Especially in the early stage of faults or when the performance deteriorates slightly, it can timely detect abnormalities through data-driven quantitative analysis, avoiding misjudgment or missed judgment caused by insufficient manual experience or unclear faults. At the same time, this method reduces the dependence on the technical level of users, and can achieve accurate evaluation of the servo system state without relying on high-level professional experience, significantly improving the objectivity and operability of diagnosis.

[0019] In addition, through real-time monitoring and intelligent analysis, it can provide continuous health state feedback for the radar servo system, helping users take preventive maintenance measures in time, avoiding further deterioration of faults or sudden shutdowns, and thus greatly improving the reliability and stability of the radar system operation. This method not only improves the maintenance efficiency of the radar system, but also extends the service life of the equipment, providing a strong guarantee for the efficient operation of the radar system in complex environments. Description of the Drawings

[0021] Figure 1 is the flowchart of the online diagnosis and evaluation method for the vibration characteristics of the radar servo system involved in the embodiment of the present invention; Figure 2 is the EMD decomposition result of the noisy simulation signal x(t) involved in the embodiment of the present invention; Figure 3 is the flowchart of the normal frequency point training of the radar servo system involved in the embodiment of the present invention; Figure 4It is a flowchart of threshold training for vibration time-domain eigenvalues (kurtosis value and root mean square value) of the radar servo system involved in the embodiments of the present invention. Specific Embodiments

[0022] The preferred mechanisms and implementation methods of the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0023] As Figures 1 to 4 shown, the embodiments of the present invention disclose an online diagnosis and evaluation method for the vibration characteristics of a radar servo system. Embodiment

[0024] An online diagnosis and evaluation method for the vibration characteristics of a radar servo system includes frequency-domain analysis, time-domain analysis, and comprehensive diagnosis and evaluation. The method flow is as Figure 1 shown.

[0025] An online diagnosis and evaluation method for the vibration characteristics of a radar servo system includes the following steps: Step 1: Receive the vibration monitoring data of the radar servo system. The vibration monitoring data includes vibration frequency-domain data and vibration time-domain data. First, preprocess the vibration frequency-domain data and vibration time-domain data, and switch to the vibration frequency-domain data or vibration time-domain data processing channel according to the data type; The early fault characteristics of the equipment are weak, the symptoms are not obvious, the fault signals are usually submerged in the strong noise background, and the operating environment is harsh and the on-site measurement conditions are limited. There will be a large number of impulsive interference signals and large-amplitude interference noises in the obtained signals. Therefore, before the diagnosis and evaluation of the system vibration characteristics, an empirical mode decomposition (EMD) noise reduction algorithm based on the intrinsic mode function (IMF) threshold is first used for data preprocessing to suppress noise, eliminate false signals and retain true signals, and improve the signal-to-noise ratio.

[0026] First, decompose the original signal into the sum of multiple IMF components and a residual component. Among them, the IMF components should meet two conditions: (1) For the entire data set, the number of zero-crossing points and extreme points must be equal or differ by at most one; (2) At any time, the average value of the envelope formed by the local maximum and minimum values should be zero.

[0027] The formula is , where L is the number of IMF components, is the i-th IMF component, and R(t) is a non-zero mean slow-varying function composed of a few extreme values, called the residual component; Then, perform a white noise test on each IMF. The formula is , where is the autocorrelation sequence calculation operator, is Corresponding autocorrelation function, take the first item of the result The absolute value is used as The difference degree from white noise. If the difference is not significant ( ), then this IMF is considered as noise and removed. Select the part of the intrinsic mode components with a large difference from Gaussian white noise ( ) for reconstruction (summation) to obtain the denoised signal. Figure 2 Is the EMD decomposition result of the noisy simulation signal x(t).

[0028] Step 2: If entering the vibration frequency domain data processing channel, first judge whether the current system speed is greater than or equal to the set threshold of 3 rpm; If the speed is less than 3 rpm, discard the vibration frequency domain data of this packet and continue to monitor the data; If the speed is greater than or equal to 3 rpm, add the vibration frequency domain data to the buffer, and judge whether the current buffer data volume is greater than or equal to the set threshold of 120; Step 3: If the buffer data volume is less than 120, continue to collect frequency domain data; If the buffer data volume is greater than or equal to 120, obtain the maximum amplitude and the frequency point information at the maximum amplitude from the parsed frequency domain waveform, and obtain the normal frequency point sets in the X / Y / Z three dimensions according to the current speed and position; Step 4: Compare and analyze with the normal frequency point set. If an anomaly occurs, add it to the anomaly set, continue to process the next packet of frequency domain data until all the buffer data is processed, and perform servo comprehensive diagnosis and evaluation; Step 5: If entering the vibration time domain data processing channel, first judge whether the current system speed is greater than or equal to the set threshold of 3 rpm; If the speed is less than 3 rpm, discard the vibration time domain data of this packet and continue to monitor the data; If the speed is greater than or equal to 3 rpm, add the vibration time domain data to the buffer, and judge whether the current buffer data volume is greater than or equal to the set threshold C; Step 6: If the buffer data volume is less than 3200, continue to collect time domain data; If the buffer data volume is greater than or equal to 3200, calculate the kurtosis values and root mean square values in the X / Y / Z three dimensions, and obtain the normal kurtosis value and root mean square value threshold information in the X / Y / Z three dimensions according to the current speed and position; Step 7: Compare and analyze with the normal kurtosis value and root mean square value thresholds. If an anomaly occurs, add it to the anomaly set, continue to process the next packet of time domain data until all the buffer data is processed, and perform servo comprehensive diagnosis and evaluation; Step 8: Enter the servo comprehensive diagnosis and evaluation process: Calculate the frequency point error. If it is an occasional anomaly, it is excluded; if it is not an occasional anomaly, report the vibration frequency domain fault. Calculate the kurtosis value and root-mean-square error. If an anomaly occurs, report the vibration time domain fault. Step 9: Based on the comprehensive analysis of the servo vibration frequency domain and time domain faults and characteristic values, judge the health level of the system.

[0029] Generally speaking, frequency domain analysis mainly includes two parts. One is to obtain the set of normal working frequency points. The process is as Figure 3 shown. Collect the frequency domain data during normal operation for training the set of normal frequency points. At this time, the training results are classified according to the rotational speed, X / Y / Z, and sensor position. After several months of training, a normal frequency point set index table is formed, and the index table is used for searching and analysis in subsequent diagnostic evaluations. The other is the diagnostic analysis of the frequency domain data during radar operation. After collecting the vibration frequency domain data, judge whether the rotational speed is greater than or equal to 3. If the current servo is not rotating or the rotational speed is too low, the vibration frequency domain data is not analyzed. When the rotational speed reaches the threshold value (frequency error tolerance limit), the data is added to the buffer area. When the data volume in the buffer area reaches 120, perform frequency domain waveform analysis, find the maximum amplitude point of the frequency domain waveform and the frequency value at that position, search for the set of normal frequency points from the index table according to the current rotational speed and position information, judge whether the current frequency value is normal, add the anomaly to the anomaly set when an anomaly occurs, and continue to process and analyze other data in the buffer area until all data in the buffer area is analyzed, end this operation, and push the anomaly set to the servo comprehensive evaluation.

[0030] Time domain analysis includes two parts. One is to obtain the time domain characteristic values of the servo during normal operation. The process is as Figure 4 shown. Collect the time domain data during normal operation for training the time domain characteristic value thresholds. At this time, the training results are classified according to the rotational speed, X / Y / Z, and sensor position. After several months of training, a normal time domain characteristic value threshold index table is formed, and the index table is used for searching and analysis in subsequent diagnostic evaluations. The time domain characteristic values include the kurtosis value and the root-mean-square value. When an anomaly occurs during the rotation of the servo, it will be intuitively reflected in these two time domain characteristic values: The formula for the kurtosis value (KU) is , where ; The formula for the root-mean-square value (RMS) is , where is the vibration amplitude value, is the number of sampling points.

[0031] Second, it is the diagnostic analysis of time-domain data during radar operation. After collecting vibration time-domain data, it is judged whether the rotational speed is greater than or equal to 3. If the current servo is not rotating or the rotational speed is too low, the vibration time-domain data will not be analyzed. When the rotational speed reaches the threshold value (amplitude error tolerance value), the data will be added to the buffer. When the data volume in the buffer reaches 3200, time-domain eigenvalue analysis will be performed. The kurtosis value and root mean square value will be calculated according to the above formula. The threshold range of the kurtosis value and root mean square value will be found from the index table based on the current rotational speed and position information to judge whether the current time-domain feature is normal. When an abnormality occurs, it will be added to the abnormal set, and the analysis of other data in the buffer will continue until all the data in the buffer is analyzed, ending this operation, and the abnormal set will be pushed to the servo comprehensive evaluation.

[0032] The comprehensive diagnostic evaluation performs comprehensive processing on the abnormal set calculated in the frequency domain and the abnormal set calculated in the time domain, which is divided into the following four situations: If both abnormal sets are empty, it means that the servo state is normal at this time, and there is no fault in the servo system, and the health level of the servo system is evaluated as normal; if the abnormal set is not empty and there are elements in the abnormal set calculated in the frequency domain or the abnormal set calculated in the time domain, the evaluation level needs to be determined according to the abnormal situation, and the error value between the abnormal frequency point and the normal frequency point will be calculated. The formula is , where is the abnormal frequency, is the normal frequency.

[0033] When the error value is greater than the threshold (frequency error tolerance value) and the occurrence times are greater than 50, it is determined as a vibration frequency domain fault. When the occurrence times are less than 50, it is determined as an occasional abnormality; According to the abnormal set of the kurtosis value and root mean square value calculated in the time domain, calculate the error value of the eigenvalues in the abnormal set. The calculation formula is , , where when , , when , ; when , , when , .

[0034] When the error value is greater than the threshold (amplitude error tolerance value), it is determined as a vibration time domain fault.

[0035] When only one dimension of the vibration frequency domain and time domain has a fault, the health level of the radar servo system is determined as inefficient.

[0036] When both the vibration frequency domain and time domain have faults, it is determined that the health level of the radar servo system is failed.

[0037] The present invention provides an online diagnosis and evaluation method for the vibration characteristics of a radar servo system. According to the monitoring data of the radar servo vibration sensor received in real time, frequency-domain eigenvalues and time-domain eigenvalues are extracted, and fault analysis and performance evaluation are carried out in two dimensions of frequency and time domain. Combining the frequency-domain and time-domain eigenvalue intervals generated during normal vibration, the health state of the current radar servo system is obtained, weakening the monitoring and diagnosis blind areas of the servo system, reducing the requirements for the high technical level of users, and improving the reliability of the operation of the radar system.

[0038] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. However, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for online diagnosis and evaluation of vibration characteristics of a radar servo system, characterized in that: The following steps are involved: Step 1: Receive vibration monitoring data of the radar servo system. The vibration monitoring data includes vibration frequency domain data and vibration time domain data. Pre-process the vibration frequency domain data and vibration time domain data first, and switch to the vibration frequency domain data or vibration time domain data processing channel according to the data type; Step 2: If entering the vibration frequency domain data processing channel, first determine whether the current system speed is greater than or equal to the set threshold A; If the rotation speed is less than A, the vibration frequency domain data of the packet is discarded and the data monitoring continues; If the rotation speed is greater than or equal to A, the vibration frequency domain data is added to the buffer area, and it is determined whether the current buffer area data volume is greater than or equal to the set threshold value B; Step 3: If the amount of data in the buffer is less than B, continue to collect frequency domain data; If the amount of data in the buffer is greater than or equal to B, the maximum amplitude and the frequency point information at the maximum amplitude are obtained from the analyzed frequency domain waveform, and the normal frequency point set in the three dimensions of X / Y / Z is obtained according to the current speed and position; Step 4: Compare and analyze with the normal frequency point set. If an abnormality occurs, add it to the abnormal set and continue to process the next packet of frequency domain data until all the buffer data is processed, and perform comprehensive servo diagnosis and evaluation; Step 5: If entering the vibration time domain data processing channel, first determine whether the current system speed is greater than or equal to the set threshold A; If the rotation speed is less than A, the vibration time domain data of the packet is discarded and the data monitoring continues; If the rotation speed is greater than or equal to A, the vibration time domain data is added to the buffer area, and it is determined whether the current buffer area data volume is greater than or equal to the set threshold C; Step 6: If the amount of data in the buffer is less than C, continue to collect time domain data; If the amount of data in the buffer is greater than or equal to C, the kurtosis value and RMS value of the three dimensions of X / Y / Z are calculated, and the normal kurtosis value and RMS value threshold information of the three dimensions of X / Y / Z are obtained according to the current speed and position; Step 7: Compare and analyze with the normal kurtosis value and RMS value threshold. If an abnormality occurs, add it to the abnormal set and continue to process the next packet of time domain data until all the buffer data is processed, and perform servo comprehensive diagnosis and evaluation.

2. The method for online diagnosis and evaluation of vibration characteristics of a radar servo system as claimed in claim 1, characterized in that: The range of threshold A is 3~6rpm, the range of threshold B is 120~150, and the range of threshold C is 3000~3200.

3. The method for online diagnosis and evaluation of vibration characteristics of a radar servo system according to claim 1 or 2, characterized in that: The threshold A set in step 2 and step 5 is 3 rpm, the threshold B set in step 2 and step 3 is 120, and the threshold C set in step 5 and step 6 is 3200.

4. The radar servo system vibration characteristic online diagnosis and evaluation method as claimed in claim 1, characterized in that: Also includes step 8: Enter the servo comprehensive diagnosis and evaluation process: Calculate the frequency point error, if it is an occasional abnormality, remove it, if it is a non-occasional abnormality, report the vibration frequency domain fault; Calculate the kurtosis value and RMS error, and if any abnormality occurs, report a vibration time domain fault.

5. The method for online diagnosis and evaluation of vibration characteristics of a radar servo system as claimed in claim 1, characterized in that: It also includes step 9: judging the health level of the system based on comprehensive analysis of servo vibration frequency domain and time domain faults and characteristic values.

6. The radar servo system vibration characteristic online diagnosis and evaluation method as claimed in claim 1, characterized in that: In step 1, the vibration time domain data and the vibration time domain data are preprocessed, and the empirical mode decomposition (EMD) denoising algorithm based on the intrinsic mode function (IMF) is used for data preprocessing.

7. The method for online diagnosis and evaluation of vibration characteristics of a radar servo system as claimed in claim 6, characterized in that: In step 1, the vibration time domain data and the vibration time domain data are preprocessed, including the following steps: S1: First, decompose the original vibration monitoring data into the sum of multiple IMF components and residual components. The formula is: , where L is the number of IMF components, is the i-th IMF component, R(t) is the residual component, which is a non-zero mean slowly varying function composed of a few extreme values; S2: Then perform Gaussian white noise test on each IMF component, the formula is ,in is the autocorrelation sequence calculation operator, for Corresponding to the autocorrelation function, take the first item of the result The absolute value is Difference from white noise; S3: If , then the difference is considered small, and the IMF component is noise and should be removed; S4: If , it is considered that the difference is large, and some inherent modal components that are larger than the Gaussian white noise are selected for reconstruction or summation to obtain the denoised data.

8. The method for online diagnosis and evaluation of vibration characteristics of a radar servo system as claimed in claim 7, characterized in that: The IMF components should meet the following conditions: For the entire vibration monitoring data set, the number of zero-crossing points and extreme value points must be equal or differ by at most one; At any time, the average value of the envelope formed by local maxima and minima should be zero.

9. The radar servo system vibration characteristic online diagnosis and evaluation method as claimed in claim 1, characterized in that: The calculation formula for the kurtosis value in step 6 is ,in ; The formula for calculating the root mean square value is: ; in is the vibration amplitude value, is the number of sampling points.

10. The method for online diagnosis and evaluation of vibration characteristics of a radar servo system as claimed in claims 5 and 9, characterized in that: According to the comprehensive analysis of servo vibration frequency domain and time domain faults and characteristic values, it can be divided into the following four situations: T1: If the abnormality sets are all empty, it means that the servo status is normal and the servo system has no faults. The health level of the radar servo system is normal. T2: If the abnormal set calculated in the frequency domain is not empty, the evaluation level is determined according to the abnormal situation, and the error value between the abnormal frequency point and the normal frequency point is calculated. The calculation formula is: ,in is the abnormal frequency, is the normal frequency; When the error value is greater than the frequency error tolerance value and the number of occurrences is greater than 50, it is determined to be a vibration frequency domain fault, and the radar servo system health level is evaluated as inefficient; When the number of occurrences is less than 50, it is considered an occasional abnormality; T3: If the anomaly set calculated in the time domain is not empty, the error value of the eigenvalue in the anomaly set is calculated based on the kurtosis value and the RMS value anomaly set calculated in the time domain. The formula is: , , among which, when hour, ,when hour, ;when hour, ,when hour, ; When the error value is greater than the amplitude error tolerance value, it is determined to be a vibration time domain fault, and the radar servo system health level is evaluated as inefficient; T4: If faults occur in both the vibration frequency domain and the time domain, the radar servo system health level is evaluated as failure.

Images