An adaptive-based magnetic suspension counter-rotating rotor misalignment vibration detection system

Abstract

The application relates to the technical field of bearing vibration detection, in particular to a magnetic suspension misalignment vibration detection system for a drag rotor based on self-adaption, which comprises a synchronous vibration signal obtained by eliminating interference in a radial vibration displacement signal through an FAPLP rotating speed frequency identification module; an LMS self-adaptive filter notch module which carries out the same frequency notch on the radial vibration displacement signal and two orthogonal signals with the same frequency as the rotating speed, so as to eliminate the main frequency component in the narrow band near the central reference frequency; and an FAPLP rotating speed 2 times frequency identification module which detects whether the output of the LMS self-adaptive filter notch module contains 2 times frequency vibration. The application only relies on the sensors and controllers equipped in the active magnetic suspension bearing to realize the misalignment vibration detection of the drag rotor, reduces the detection cost, and improves the detection speed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bearing vibration detection, and particularly relates to a self-adaptive magnetic suspension pair-rotor misalignment vibration detection system. BACKGROUND

[0002] With the rapid development of industrial modernization, China is gradually moving from a manufacturing power to a manufacturing superpower. Compared with western developed countries, China's manufactured mechanical and electrical products generally have the characteristics of low technical content and lack of independent intellectual property rights. Therefore, China proposes the "Made in China 2025" national strategy to vigorously promote advanced manufacturing. As a basic technology in the field of industrial equipment, supporting technology is widely used in chemical industry, machinery, aviation and other industries. Among them, bearings, as the key supporting components of rotors to bear loads, are increasingly demanding in terms of meeting conditions, such as high speed, high temperature, large load, and complex environment.

[0003] The rotor is the core component of the rotating machinery and is also the main part of the fault. With various unstable factors, many rotors of rotating machinery will inevitably produce some faults, such as rotor mass imbalance, shaft bending, rotor dynamic and static rubbing, loose connection near the rotor, shaft misalignment, oil film whirling and oil film oscillation, etc. Once a fault occurs, it may cause some functions of the mechanical equipment to weaken or be lost, or even cause a chain reaction, leading to the failure of the entire equipment to work normally or even paralysis, and even cause personnel casualties. Once a mechanical equipment accident caused by rotor failure occurs, it will cause great loss to the national economy.

[0004] Traditional magnetic suspension bearing vibration detection and control research focuses on the same frequency vibration caused by rotor mass imbalance. However, with the rapid development of rotating machinery, the application range of pair-rotor with coupling is also increasingly widespread. According to relevant statistics, most of the rotor misalignment problems are caused by couplings. This kind of fault has less impact in small magnetic suspension bearings with low power and low load, but in large magnetic suspension bearings with high speed and high power, the misalignment caused fault is obviously increased. Therefore, solving the misalignment vibration problem of magnetic suspension pair-rotor is not only beneficial to the stable operation of rotating machinery and the extension of service life, but also has far-reaching significance for the development of magnetic suspension pair-rotor. SUMMARY

[0005] In view of the defects of the prior art, the technical scheme adopted by the present application is: a self-adaptive magnetic suspension counter-rotor misalignment vibration detection system based on FAPLP, comprising: an FAPLP rotating speed frequency identification module, an LMS self-adaptive filter notch module and an FAPLP rotating speed 2 times frequency identification module, the output end of the FAPLP rotating speed frequency identification module is connected with the input end of the LMS self-adaptive filter notch module, and the output end of the LMS self-adaptive filter notch module is connected with the input end of the FAPLP rotating speed 2 times frequency identification module; the synchronous vibration signal in the radial vibration displacement signal is eliminated through the FAPLP rotating speed frequency identification module; the radial vibration displacement signal and two orthogonal signals with the same frequency as the rotating speed are subjected to the same frequency notch through the LMS self-adaptive filter notch module, so that the main frequency component in the narrow band near the center reference frequency is eliminated; whether the output of the LMS self-adaptive filter notch module contains 2 times frequency vibration is detected through the FAPLP rotating speed 2 times frequency identification module.

[0006] Further, the radial vibration displacement signal is collected through the eddy current displacement sensor.

[0007] Further, the FAPLP rotating speed frequency identification module comprises:

[0008] The input is the radial vibration displacement signal u(n), and a high-order linear prediction signal is established:

[0009]

[0010] Wherein, c(i) is a linear prediction coefficient, and M is the order of the filter;

[0011] The expression of the prediction error is defined as:

[0012]

[0013] Wherein, U n-1 and C n-1 are the input vector and the weight vector;

[0014] According to the minimum mean square error criterion, the optimal linear prediction coefficient vector is:

[0015]

[0016] According to the M-order error equation, M roots are obtained:

[0017]

[0018] The estimated frequency is calculated as:

[0019]

[0020] The estimated value of the signal frequency is solved by using the average phase method:

[0021]

[0022] wherein, F s is the sampling frequency.

[0023] Further, the LMS adaptive filter trap module comprises:

[0024] The reference input X(n) is set as:

[0025]

[0026] wherein, Ω0 is the rotating speed, T is the sampling time, is the initial phase; the LMS algorithm is adopted to carry out adaptive filtering, and the filter weight value is calculated, and the formula is:

[0027]

[0028] The beneficial effects of the present application are:

[0029] 1. The present application can obtain the rotor rotating speed information without prior knowledge, i.e. without additional detection equipment, and only relies on the sensor and controller equipped in the active magnetic bearing to realize the detection of the misalignment vibration of the drag rotor, which not only reduces the detection cost, but also improves the detection speed.

[0030] 2. The present application adopts the LMS adaptive filter to carry out real-time estimation and update on the coefficient vector of the prediction model, so as to ensure the real-time performance of the frequency estimation.

[0031] 3. The average phase method is adopted to process the prediction coefficient vector to avoid the operation of solving the Mth equation by the ALP algorithm, so as to reduce the complexity of the algorithm, and the algorithm has strong robustness, fast calculation speed and strong stability. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 is the system block diagram of the misalignment vibration detection system of the drag rotor of the active magnetic suspension based on the present application;

[0033] Figure 2 is the principle block diagram of the FAPLP rotating speed frequency identification;

[0034] Figure 3 is the principle block diagram of the LMS adaptive filter trap;

[0035] Figure 4 is the weight adaptive iteration process diagram of the LMS adaptive algorithm;

[0036] Figure 5 is the simulation experimental result diagram. DETAILED DESCRIPTION

[0037] The application will be further described below in connection with the accompanying drawings and examples, which are simplified schematic diagrams only showing the basic structure of the application in a schematic manner, and thus only show the components related to the application.

[0038] As Figure 1 As shown in the system block diagram, the adaptive-based magnetic suspension counter-rotating rotor misalignment vibration detection system comprises an FAPLP rotor speed frequency identification module, an LMS adaptive filter notch module and an FAPLP rotor speed 2 times frequency identification module, the output end of the FAPLP rotor speed frequency identification module is connected with the input end of the ALMS adaptive filter notch module, and the output end of the LMS adaptive filter notch module is connected with the input end of the FAPLP rotor speed 2 times frequency identification module.

[0039] The displacement sensor detects the radial vibration displacement signal of the active magnetic suspension counter-rotating rotor, and sends the radial vibration displacement signal as input to the FAPLP rotor speed frequency identification module of the controller, so that the frequency identification can be realized under the condition of high-frequency noise interference, and no additional sensor and little prior knowledge are needed.

[0040] The radial vibration displacement signal of the magnetic suspension counter-rotating rotor when misalignment fault occurs is obtained by using the eddy current displacement sensor, the radial vibration displacement signal and two orthogonal signals with the same frequency as the rotor speed are input to the LMS adaptive filter notch module for same-frequency notch, so as to eliminate the main frequency component in the narrow band near the center reference frequency, since the adaptive notch filter only has one parameter to be estimated, the structure is simple, the control bandwidth, zero depth are large, and the interference frequency and phase can be accurately adaptively tracked, and the stability is ensured by limiting the pole within the unit circle.

[0041] The signal after the notch rotor speed same-frequency is input to the FAPLP rotor speed 2 times frequency identification module, when the magnetic suspension bearing counter-rotating rotor misalignment fault occurs, the radial vibration not only contains the rotor speed same-frequency vibration, but also contains the 2 times frequency vibration.

[0042] As Figure 2 The FAPLP rotor speed frequency identification module is shown, the input is the radial vibration displacement signal u(n), and a high-order linear prediction signal is established:

[0043]

[0044] Wherein, c(i) is the linear prediction coefficient, and M is the order of the filter; according to formula (1), the expression of the prediction error is defined as:

[0045]

[0046] Wherein, U n-1 and C n-1is the input vector and w is the weight vector.

[0047] According to the minimum mean square error criterion, the optimal linear prediction coefficient vector is:

[0048]

[0049] The expression of the Z-domain prediction error is:

[0050]

[0051] According to the M-order error equation, M roots are:

[0052]

[0053] Q O The root closest to the unit circle of the error equation defines the value of where r0 is the radius of the unit circle and θ0 is the estimated phase information; when the sampling frequency is F s , the estimated frequency is:

[0054]

[0055] Since the calculation complexity of solving the error prediction is very high, it is not suitable for real-time application. In order to improve the solving speed, the average phase method is proposed instead of solving the error equation; at the same time, the calculation speed is greatly improved with little accuracy. In practice, due to noise interference, there will be a certain error in each prediction coefficient. The frequency estimation obtained by the first-order structure will have a large error. Therefore, the average phase is used to reduce the noise influence and improve the accuracy.

[0056]

[0057] The estimated value of the signal frequency is:

[0058]

[0059] For the 2 times frequency vibration caused by the rotor misalignment fault, as long as it is estimated whether the notch signal after the LMS notch filter module contains the rotational speed 2 times frequency, so as to judge whether the misalignment vibration fault occurs, its essence is the same, so the FAPLP rotational speed 2 times frequency identification module also identifies the rotational speed 2 times frequency based on the FAPLP rotational speed frequency identification module.

[0060] As Figure 3The shown is the principle block diagram of LMS adaptive filter notch, the radial vibration displacement signal is sent to the original input end; the reference input end is two orthogonal signals with the same frequency as the rotating speed, the purpose is to obtain two weight values w1(n) and w2(n), so that the amplitude and phase of the combined sine wave can be the same as those of the interference component in the original input; the weighted output of the reference input is y(n), which is the estimation of the radial vibration displacement main frequency. Thus, the displacement signal eliminating the same frequency vibration is obtained by subtracting the estimated signal from the sampling signal containing 2 times the frequency.

[0061] The reference input X(n) is set as:

[0062]

[0063] Wherein, Ω0 is the rotating speed, T is the sampling time, is the initial phase; the LMS algorithm is used for adaptive filtering, as shown in the figure Figure 4 The correction process of the filter weight value is as follows:

[0064]

[0065] In the formula, w(n) is the weight value of adaptive iteration, μ is the step length of adaptive iteration, and X(n) is the reference input.

[0066] Figure 5 The shown is the detection result of the misalignment fault of the magnetic suspension counter-rotor; wherein, Figure 5 (a) is the radial vibration displacement signal detected by the displacement sensor, which contains the same frequency vibration caused by mass imbalance, 2 times the frequency vibration caused by misalignment fault, and high frequency noise and other disturbances.

[0067] Figure 5 (b) is the LMS adaptive filter notch module, which can realize accurate identification of the rotating speed frequency under the condition that the signal-to-noise ratio is high;

[0068] Figure 5 (c) is the LMS adaptive filter notch module for eliminating the rotating speed signal with the same frequency;

[0069] Figure 5 (d) uses the detection result of the FAPLP rotating speed frequency identification module, which is the frequency result and the amplitude result respectively; when the 2 times the frequency signal of the rotating speed frequency appears in the detection result, it can be considered that the counter-rotor has misalignment fault, and the greater the amplitude signal is, the greater the misalignment of the counter-rotor is.

[0070] With the above ideal embodiments according to the present application as the inspiration, through the above description, relevant staff can make various changes and modifications without deviating from the technical idea of the present application. The technical scope of the present application is not limited to the content of the specification, and must be determined according to the scope of the claims.

Claims

1. A self-adaptive magnetic suspension misalignment vibration detection system for a contra-rotating rotor, characterized in that, The FAPLP rotating speed frequency identification module, the LMS adaptive filter notch module and the FAPLP rotating speed 2 times frequency identification module, the output end of the FAPLP rotating speed frequency identification module is connected with the input end of the LMS adaptive filter notch module, the output end of the LMS adaptive filter notch module is connected with the input end of the FAPLP rotating speed 2 times frequency identification module; the synchronous vibration signal in the radial vibration displacement signal is eliminated through the FAPLP rotating speed frequency identification module; the radial vibration displacement signal and two orthogonal signals with the same frequency as the rotating speed are subjected to the same frequency notch through the LMS adaptive filter notch module, so that the main frequency component in the narrow band near the central reference frequency is eliminated; whether the output of the LMS adaptive filter notch module contains 2 times frequency vibration is detected through the FAPLP rotating speed 2 times frequency identification module; The FAPLP rotating speed frequency identification module comprises: The expression for defining the prediction error is: Let the input be a radial vibration displacement signal u ( n ), a high-order linear prediction signal is established: ⑴ wherein c i ) is a linear prediction coefficient, M is the order of the filter;​ According to the least mean square error criterion, the optimal linear prediction coefficient vector is: ⑵ wherein, U n-1 and C n-1 are input vectors and weight vectors; The estimated frequency is calculated as: ⑶ According to M The order error equation is obtained M The roots are: （5） The estimated value of the signal frequency is solved by using the average phase method as: ⑹ The LMS adaptive filter notch module comprises: ⑻ wherein F s is the sampling frequency, ; The reference input X(n) is set as: The radial vibration displacement signal is collected by the eddy current displacement sensor. ⑼ wherein, is the rotational speed, T is the sampling time, is the initial phase; the LMS algorithm is used for adaptive filtering, and the filter weight value is calculated according to the formula: ⑽ where w(n) is the weight of the adaptive iteration, is the step size of the adaptive iteration, and X(n) is the reference input.

2. The adaptive based magnetic levitation misalignment vibration detection system for a counter-rotating rotor pair according to claim 1, wherein ​

Images