Antenna servo drive state monitoring and fault diagnosis method
By combining stator current signal monitoring with wavelet transform and singular value decomposition, the feasibility problem of servo drive status monitoring and fault diagnosis in aerospace telemetry and control equipment has been solved, achieving efficient and accurate fault diagnosis, which is particularly suitable for automated equipment in harsh environments.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA XIAN SATELLITE CONTROL CENT
- Filing Date
- 2023-02-22
- Publication Date
- 2026-05-29
AI Technical Summary
The feasibility of servo drive status monitoring and fault diagnosis for existing aerospace telemetry and control equipment is difficult. The diagnostic algorithms are complex and have low effectiveness, and the vibration signal acquisition is costly and complex.
Servo motor fault diagnosis is performed using stator current signals. The stator three-phase current signals are collected by Hall current sensors, and after noise reduction preprocessing, frequency domain features and relative normalized amplitude features are extracted to establish the correspondence between different states of the servo motor. Fault diagnosis is then performed by combining wavelet transform and singular value decomposition.
It enables accurate monitoring and fault diagnosis of servo drive status, reduces monitoring costs, improves diagnostic effectiveness, and is suitable for automated operation equipment in harsh environments.
Smart Images
Figure CN116184193B_ABST
Abstract
Description
Technical Field
[0001] This invention patent belongs to the field of spacecraft measurement, control and application, and relates to a method for antenna servo drive status monitoring and fault diagnosis. Background Technology
[0002] The increasingly frequent space launch missions place higher demands on the stable operation of ground equipment such as radar and antennas. Servo motors play an important role in radar and antennas, and their stable operation determines the antenna's operating performance and tracking accuracy. However, some ground equipment operates under harsh conditions such as high temperature, high pressure, and humidity for a long time, which can easily cause structural damage to the stator, rotor, and bearings of servo motors in the azimuth axis, pitch axis, and third axis. Accurate and rapid fault diagnosis methods are needed to identify faults and deal with them in a timely manner.
[0003] Currently, the more mature motor fault diagnosis methods mainly include three aspects: (1) analytical model-based methods, which require the establishment of a relatively accurate motor model and have certain difficulties for complex nonlinear systems; (2) knowledge-based methods, which have the problem of imperfect algorithm universality; (3) signal processing-based methods, which can directly use vibration signals and electrical quantities during the operation of servo motors as feature monitoring parameters and have strong applicability; however, the vibration signals are collected by embedded sensors, which increases the monitoring cost and complexity.
[0004] To address the aforementioned issues, this application introduces stator current signals for servo motor fault diagnosis. Stator current signals offer advantages such as convenient acquisition, high sensitivity, strong intuitiveness of time-domain characteristics, and good stability. Using stator current signals as fault diagnosis feature parameters can effectively extract weak fault feature components that are submerged by fundamental frequency components, channel noise, eccentric harmonics, etc. Summary of the Invention
[0005] The purpose of this invention is to provide a method for monitoring the status of antenna servo drives and diagnosing faults, which solves the problems of difficult implementation, complex diagnostic algorithms, and low effectiveness of existing servo drive status monitoring and fault diagnosis methods for aerospace telemetry and control equipment.
[0006] The technical solution adopted in this invention is an antenna servo drive status monitoring and fault diagnosis method, the specific steps of which are as follows:
[0007] Step 1: Collect the stator three-phase current signals of the antenna servo motor under normal and fault conditions using Hall current sensors. Perform noise reduction preprocessing on the collected stator three-phase current signals to obtain the reconstructed three-phase current signals under each condition.
[0008] Step 2: Extract frequency domain features and relative normalized amplitude features from the reconstructed three-phase current signals under different states, and establish the correspondence between the normalized amplitude of the three-phase current and its frequency domain features under each state of the servo motor.
[0009] Step 3: Perform servo motor fault diagnosis based on the correspondence between the relative normalized amplitude characteristics of the three-phase current of the servo motor under test and its frequency domain characteristics.
[0010] The invention is further characterized in that,
[0011] The current signal noise reduction preprocessing process is as follows:
[0012] Step 1.1: Define any current signal as i s (t), where t represents the sampling time, for the current signal i s (t) is decomposed into two layers of wavelets to initially remove high-frequency noise components from the signal and obtain the current signal i1(t);
[0013] Step 1.2: Perform singular value decomposition (SVD) on the current signal i1(t) to reduce noise. Construct an m×n Hankel matrix X, and perform SVD on the matrix X to obtain X = USV. T , Σ=diag(λ1,λ2,...λ k ..,λ r ), where U is a 1000×1000 unitary matrix and V is a 1001×1001 unitary matrix. Σ is the sequence of singular values of matrix X, λ1,λ2,...λ k ..,λ r These are the singular values in the singular value sequence, λ1, λ2, ..., λ. k ..,λ r >0, r is the total number of singular values in the singular value sequence, and k is the number of singular value points that satisfy the Wright criterion;
[0014] Step 1.3: Calculate the arithmetic mean of the singular value sequence Σ. And the standard deviation σ, the specific calculation process is as follows:
[0015]
[0016]
[0017] Based on the singular value sequence Σ and the arithmetic mean Calculate the sequence residual λ b :
[0018]
[0019] Where, λ bLet λ(i) be the sequence residual, λ(i) (i = 1, 2... r) be the singular points that reflect the useful signal, σ be the standard deviation, and r be the total number of singular values in the singular value sequence.
[0020] Step 1.4: According to the Wright criterion, the residual λ b Compare each one with 3 times the standard deviation 3σ, if |λ b If |i|≥3σ, then λ(i) are the singular points that reflect the useful signal. These k singular points that satisfy the Wright criterion are retained, and those that do not satisfy |λ| are discarded. b (i) Setting the (rk) smaller singular values required by |≥3σ to 0 yields the reconstructed Hankel matrix X. s , where i = 1, 2, ..., k, and k is the number of singular points that satisfy the Wright criterion;
[0021] Step 1.5: Based on the reconstructed Hankel matrix X s The correspondence between the elements in the sequence and the current sequence yields the reconstructed current signal I(t) after noise removal;
[0022] Step 1.6: Repeat steps 1.1-1.5 to complete the reconstruction of all three-phase current signals acquired in step 1.
[0023] The specific steps for step 2 are as follows:
[0024] Step 2.1: Differentiate any reconstructed current signal I(t) and find the maximum point I of I(t). maxp Minimum point I minq and its corresponding time τ maxp τ minq Where p = 1, 2…m1, q = 1, 2…m2, m1 and m2 are the number of maxima and minima, respectively; Step 2.2: Use the cubic spline interpolation function to reconstruct the extreme points (It, Iq, ..., m2) of the current signal I(t). maxp ,τ maxp ), (I minq ,τ minq Fitting is performed to obtain the upper and lower envelope curves of the reconstructed current signal I(t), thereby filtering out the power grid frequency component of the current signal;
[0025] Step 2.3: Perform spectral analysis on the envelope curve of the reconstructed current signal I(t) using Fast Fourier Transform to obtain the frequency domain characteristics of the fault characteristic frequency;
[0026] Step 2.4: Use the absolute value operator to demodulate the envelope curve of the reconstructed current signal I(t), extract the amplitude characteristics of the envelope curve, and then its relative normalized amplitude A(x(t)) is:
[0027] A(x(t))=|x(t)| (4)
[0028] Where x(t) is the modulation signal;
[0029] Step 2.5: Repeat steps 2.1-2.4 to extract the frequency domain features and relative normalized amplitude features of all reconstructed current signals obtained in step 1.6. Obtain the frequency domain features and relative normalized amplitude of each reconstructed current signal, and establish the correspondence between the normalized amplitude of the three-phase current and its frequency domain features under each state of the servo motor.
[0030] The servo motor faults in step 1 include stator faults, rotor faults, bearing outer ring faults, bearing inner ring faults, bearing rolling element faults, and bearing cage faults.
[0031] The specific correspondence between the various states of the servo motor and the frequency domain characteristics and relative normalized amplitude characteristics of the three-phase current are as follows:
[0032]
[0033]
[0034] Among them, A a A b A c These represent the relative normalized amplitudes of the three-phase current envelope curves of the servo motor under the same state, f. s Here, s is the power frequency component, s is the motor slip, and f is the motor slip. o f r f b and f c These are the characteristic frequencies of vibration signals from the bearing outer ring, inner ring, rolling elements, and cage, respectively, where n' = 1, 2, 3…, f cage To maintain the rotational angular frequency of the cage.
[0035] The m×n order Hankel matrix X is as follows:
[0036]
[0037] Where m = 1000, n = 1001.
[0038] The beneficial effects of this invention are:
[0039] (1) This invention establishes various state models of servo drive, proposes a stator current signal noise suppression method combining wavelet transform and SVD, and proposes an envelope curve demodulation algorithm to effectively extract the frequency domain features corresponding to the normal state of antenna servo drive, stator fault, rotor fault and bearing fault, so as to realize the servo drive state monitoring and fault diagnosis of measurement and control equipment.
[0040] (2) The method of the present invention effectively solves the problems of inconvenient vibration signal acquisition and numerous types of characteristic parameter monitoring, and provides important guarantee for the stable operation of existing ground measurement and control equipment. It is especially suitable for automated operation equipment in harsh environments where "there are people on duty but no one is on duty". Attached Figure Description
[0041] Figure 1 This is a block diagram of the overall structure design of the servo drive status monitoring and fault diagnosis system used in the antenna servo drive status monitoring and fault diagnosis method of the present invention.
[0042] Figure 2 This is a flowchart of the stator current signal envelope demodulation analysis in the antenna servo drive status monitoring and fault diagnosis method of the present invention;
[0043] Figure 3 This is a time-domain waveform diagram of the stator current envelope curve for a bearing inner ring fault in the antenna servo drive status monitoring and fault diagnosis method of the present invention.
[0044] Figure 4 This is the frequency domain characteristic diagram of the stator current upper envelope curve in the antenna servo drive status monitoring and fault diagnosis method of the present invention;
[0045] Figure 5 This is a comparison chart of the relative normalized amplitudes of the three-phase current envelope curves of the bearing inner ring in the antenna servo drive status monitoring and fault diagnosis method of the present invention. Detailed Implementation
[0046] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.
[0047] To make the objectives, technical solutions, and advantages of the present invention clearer, the specific steps of steps 1 and 2 of the present invention are described using the bearing inner ring failure mode as an example, and the present invention will be further described in detail.
[0048] Step 1: Collect the three-phase current signal of the antenna servo motor bearing inner ring under fault conditions using a Hall current sensor. Perform noise reduction preprocessing on the collected three-phase current signals to obtain the reconstructed three-phase current signal.
[0049] Figure 1 The overall structure of the servo drive status monitoring and fault diagnosis system adopted in this invention is designed based on the structure of ground measurement and control equipment and actual use. The system mainly consists of a signal acquisition module, a data processing module and a monitoring and diagnosis module.
[0050] Three Hall current sensors are connected to the three-phase stator windings a, b, and c of the antenna servo motor respectively to collect the three-phase current signal under the bearing inner ring fault mode. At this time, the three-phase current signal contains noise interference such as system noise, mechanical noise, and electromagnetic noise. After preprocessing by the data processing module, the noise interference is effectively suppressed and the useful signal is amplified.
[0051] The noise interference suppression steps for any current signal are as follows:
[0052] Step 1.1: Employ wavelet threshold denoising, selecting the sym8 wavelet function, to denoise the acquired noisy stator current signal i. s (t) is decomposed into two levels of wavelet decomposition to initially remove high-frequency noise components from the signal, resulting in i1(t);
[0053] Step 1.2: Perform singular value decomposition (SVD) on the current signal i1(t) to reduce noise. Construct an m×n Hankel matrix X, and perform singular value decomposition on matrix X to obtain X = USV. T Where U is a 1000×1000 unitary matrix and V is a 1001×1001 unitary matrix. Σ=diag(λ1,λ2,...λ k ..,λ r ), λ1,λ2,...λ k ..,λ r >0, Σ is the sequence of singular values of matrix X, λ1,λ2,...λ k ..,λ r These are the singular values in the singular value sequence;
[0054] The m×n order Hankel matrix X is as follows:
[0055]
[0056] Where m = 1000, n = 1001;
[0057] Step 1.3: Calculate the arithmetic mean of the singular value sequence Σ. And the standard deviation σ, the specific calculation process is as follows:
[0058]
[0059]
[0060] Based on the singular value sequence Σ and the arithmetic mean Calculate the sequence residual λ b :
[0061]
[0062] Where, λb Let λ(i) be the sequence residual, λ(i) (i = 1, 2... r) be the singular points that reflect the useful signal, σ be the standard deviation, and r be the total number of singular values in the singular value sequence.
[0063] Step 1.4: According to the Wright criterion, the residual λ b Compare each one with 3 times the standard deviation 3σ, if |λ b If |i|≥3σ, then λ(i) are the singular points that reflect the useful signal. These k singular points that satisfy the Wright criterion are retained, and those that do not satisfy |λ| are discarded. b (i) Set the (rk) smaller singular values required by |≥3σ to 0 to obtain the reconstructed Hankel matrix X. s , where i = 1, 2, ..., k, and k is the number of singular points that satisfy the Wright criterion;
[0064] Step 1.5: Based on the reconstructed Hankel matrix X s The correspondence between the elements in the current sequence and the current sequence is used to obtain the reconstructed current signal I(t) after removing noise;
[0065] Step 1.6: Repeat steps 1.1-1.5 to complete the reconstruction of the three-phase current signal of the stator winding under the bearing inner ring fault mode.
[0066] Step 2: Extract the frequency domain characteristics and envelope curve amplitude characteristics of the reconstructed three-phase current signal under the bearing inner ring fault mode, and establish the correspondence between the normalized amplitude of the three-phase current and its frequency domain characteristics under each state of the servo motor.
[0067] The specific process is as follows: Figure 2 As shown, the specific steps are as follows:
[0068] Step 2.1: Differentiate any reconstructed current signal I(t) and find the maximum point I of I(t). maxp Minimum point I minq and its corresponding time τ maxp τ minq Where p = 1, 2…m1, q = 1, 2…m2, and m1 and m2 are the number of local maxima and local minima, respectively;
[0069] Step 2.2: Use the cubic spline interpolation function to reconstruct the extreme points (It, It, It) of the current signal I(t). maxp ,τ maxp ), (I minq ,τ minq Fitting is performed to obtain the upper and lower envelope curves of the reconstructed current signal I(t), thereby filtering out the power grid frequency component of the current signal. The time-domain waveform of the bearing inner ring fault current envelope curve is shown below. Figure 3 As shown;
[0070] Step 2.3: Perform spectral analysis on the envelope curve of the reconstructed current signal I(t) using Fast Fourier Transform (FFT) to obtain the frequency domain characteristics reflecting the fault characteristic frequencies. The time domain waveform of the bearing inner ring fault phase current envelope curve is shown below. Figure 3 As shown, the frequency domain characteristics of the current upper envelope curve are as follows: Figure 4 As shown;
[0071] Step 2.4: Demodulate the envelope curve of the reconstructed current signal I(t) using the absolute value operator, extract the amplitude characteristics of the envelope curve, and then its relative normalized amplitude A(x(t)) is:
[0072] A(x(t))=|x(t)| (4)
[0073] Where x(t) is the modulation signal;
[0074] Step 2.5: Repeat steps 2.1-2.4 to complete the frequency domain characteristics and relative normalized amplitude characteristics of the three-phase reconstructed current signals. Obtain the correspondence between the frequency domain characteristics and relative normalized amplitude of each reconstructed current signal. Establish the correspondence between the relative normalized amplitude characteristics of the three-phase currents and their correspondence with the frequency domain characteristics under the bearing inner ring fault mode, as follows: Figure 5 As shown.
[0075] Repeat steps 1 and 2 to establish the correspondence between the normalized amplitude of the three-phase current and its frequency domain characteristics under each state of the servo motor. The specific correspondence is as follows:
[0076]
[0077]
[0078] Among them, A a A b A c These represent the relative normalized amplitudes of the three-phase current envelope curves of the servo motor under the same state, f. s Here, s is the power frequency component, s is the motor slip, and f is the motor slip. o f r f b and f c These are the characteristic frequencies of vibration signals from the bearing outer ring, inner ring, rolling elements, and cage, respectively, where n' = 1, 2, 3…, f cage To maintain the rotational angular frequency of the cage.
[0079] The specific fault diagnosis process is as follows:
[0080] Three Hall current sensors are connected to the three-phase stator windings of the antenna servo motor respectively to collect the three-phase current signals of the stator windings of the servo motor. According to the method of the present invention, the correspondence between the current three-phase current of the servo motor and the normalized amplitude and its correspondence with the frequency domain characteristics are obtained.
[0081] If the relative normalized amplitudes of the three-phase currents are the same and equal to 1, and the upper and lower envelope curves of the current signal have no sideband components in the frequency domain, then the servo motor is operating normally. If the relative normalized amplitudes of the three-phase currents are all different, and the upper and lower envelope curves of the current signal have no sideband components in the frequency domain, then the servo motor stator is faulty. If the relative normalized amplitudes of the three-phase currents are the same and equal to 1, and the upper and lower envelope curves of the current signal have prominent sideband components in the frequency domain, with a frequency value of (1±2s)f... s If the three-phase currents have the same relative normalized amplitude and are equal to 1, the upper and lower envelope curves of the current signal exhibit prominent sideband components in the frequency domain, and the frequency value is |f s ±n'f o If the three-phase currents have the same relative normalized amplitude and are equal to 1, the sideband components of the upper and lower envelope curves of the current signal will be prominent. s ±f r |、|f s ±f r ±n'f i If the three-phase currents have the same relative normalized amplitude and are equal to 1, the upper and lower envelope curves of the current signal exhibit prominent sideband components in the frequency domain, and the frequency value is |f. s ±f cage |、|f s ±f cage ±n'f b If the three-phase currents have the same relative normalized amplitude and are equal to 1, the sideband components of the upper and lower envelope curves of the current signal are prominent in the frequency domain, and the frequency value is |f. s ±n'f c If the bearing cage fails, then the bearing cage is faulty.
[0082] Based on engineering practice verification of a certain type of ground measurement and control equipment, the above steps can effectively realize the status monitoring and fault diagnosis of antenna azimuth and elevation axis servo drives.
Claims
1. A method for monitoring the status and diagnosing faults of an antenna servo drive, characterized in that, The specific steps are as follows: Step 1: Collect the stator three-phase current signals of the antenna servo motor under normal and fault conditions using Hall current sensors. Perform noise reduction preprocessing on the collected stator three-phase current signals to obtain the reconstructed three-phase current signals under each condition. Step 2: Extract frequency domain features and relative normalized amplitude features from the reconstructed three-phase current signals under different states, and establish the correspondence between the normalized amplitude of the three-phase current and its frequency domain features under each state of the servo motor. Step 3: Perform servo motor fault diagnosis based on the correspondence between the relative normalized amplitude characteristics of the three-phase current of the servo motor under test and its frequency domain characteristics.
2. The antenna servo drive status monitoring and fault diagnosis method according to claim 1, characterized in that, The current signal noise reduction preprocessing process is as follows: Step 1.1: Define any current signal as... i s ( t ), t Indicates the sampling time for the current signal. i s ( t Two-level wavelet decomposition is performed to initially remove high-frequency noise components from the signal, yielding the current signal. i 1( t ); Step 1.2: Analyze the current signal. i 1( t Perform singular value decomposition for noise reduction and construct... Hankel matrix of order X For the matrix X Perform singular value decomposition to obtain X = USV T , ,in, U for unitary matrix of order, V for unitary matrix of order, , For matrix X The sequence of singular values, These are the singular values in the singular value sequence. , r This represents the total number of singular values in the singular value sequence. k The number of singular points to satisfy the Wright criterion; Step 1.3: Calculate the singular value sequence arithmetic mean and standard deviation The specific calculation process is as follows: （1） （2） Based on singular value sequences and arithmetic mean Calculate the sequence residuals : （3） in, For sequence residuals, To reflect the singular points of the useful signal, Standard deviation, r This represents the total number of singular values in the singular value sequence. Step 1.4: According to the Wright criterion, convert the residuals... One by one and three times the standard deviation Compare, if ,but To reflect the singular points of the useful signal, retain this k A singular value that satisfies the Wright criterion will not satisfy the Wright criterion. Required ( r - k By setting the smaller singular values to 0, the reconstructed Hankel matrix is obtained. X s ,in, , k The number of singular points to satisfy the Wright criterion; Step 1.5: Based on the reconstructed Hankel matrix X s The correspondence between the elements and the current sequence yields the reconstructed current signal after noise removal. I ( t ); Step 1.6: Repeat steps 1.1-1.5 to complete the reconstruction of all three-phase current signals acquired in step 1.
3. The antenna servo drive status monitoring and fault diagnosis method according to claim 1, characterized in that, The specific steps for step 2 are as follows: Step 2.1: For any reconstructed current signal I ( t Take the derivative and find I ( t The maximum point of ) I maxp Minimum point I minq and their corresponding times τ maxp , τ minq ,in, p =1,2… m 1, q =1,2… m 2, m 1. m 2 represents the number of maximum and minimum points, respectively; Step 2.2: Reconstruct the current signal using a cubic spline interpolation function. I ( t extreme points of ) I maxp , τ maxp ), ( I minq , τ minq The reconstructed current signal is obtained by fitting the data. I ( t The upper and lower envelope curves of the current signal are used to filter out the power grid frequency component of the current signal. Step 2.3: Reconstruct the current signal using Fast Fourier Transform. I ( t Spectral analysis was performed on the envelope curve to obtain the frequency domain characteristics of the fault characteristic frequencies; Step 2.4: Reconstruct the current signal using the absolute value operator. I ( t Demodulate the envelope curve and extract its amplitude characteristics; then its relative normalized amplitude is obtained. for; （4） in, x ( t () is the modulated signal; Step 2.5: Repeat steps 2.1-2.4 to extract the frequency domain features and relative normalized amplitude features of all reconstructed current signals obtained in step 1.
6. Obtain the frequency domain features and relative normalized amplitude of each reconstructed current signal, and establish the correspondence between the normalized amplitude of the three-phase current and its frequency domain features under each state of the servo motor.
4. The antenna servo drive status monitoring and fault diagnosis method according to claim 1, characterized in that, The servo motor faults mentioned in step 1 include stator faults, rotor faults, bearing outer ring faults, bearing inner ring faults, bearing rolling element faults, and bearing cage faults.
5. The antenna servo drive status monitoring and fault diagnosis method according to claim 1, characterized in that, The specific correspondence between the various states of the servo motor and the frequency domain characteristics and relative normalized amplitude characteristics of the three-phase current is as follows: If the relative normalized amplitudes of the three-phase currents are the same and equal to 1, and the upper and lower envelope curves of the current signal have no sideband components in the frequency domain, then the servo motor is operating normally. If the relative normalized amplitudes of the three-phase currents are all different, and the upper and lower envelope curves of the current signal have no sideband components in the frequency domain, then the servo motor stator is faulty. If the relative normalized amplitudes of the three-phase currents are the same and equal to 1, and the upper and lower envelope curves of the current signal have prominent sideband components in the frequency domain, with a frequency value of [missing value], then the servo motor is operating normally. If the three-phase currents have the same relative normalized amplitude and are equal to 1, the upper and lower envelope curves of the current signal exhibit prominent sideband components in the frequency domain, and the frequency value is... If the bearing outer ring is faulty, then the three-phase current has the same relative normalized amplitude and is 1; if the upper and lower envelope curves of the current signal have prominent sideband components. If the bearing inner ring is faulty, then the three-phase currents have the same relative normalized amplitude and are equal to 1. The upper and lower envelope curves of the current signal exhibit prominent sideband components in the frequency domain, and their frequency values are... If the three-phase currents have the same relative normalized amplitude and are equal to 1, the sideband components of the upper and lower envelope curves of the current signal are prominent in the frequency domain, and the frequency value is... If so, the bearing cage is faulty; in, f s The power frequency component is the mains frequency component. s This refers to the motor slip rate. f o , f r , f b and f c These are the characteristic frequencies of vibration signals from the bearing's outer ring, inner ring, rolling elements, and cage, respectively. =1,2,3…, To maintain the rotational angular frequency of the cage.
6. The antenna servo drive status monitoring and fault diagnosis method according to claim 2, characterized in that, The Hankel matrix of order X as follows: （5） in, m =1000, n =1001.