Rotor collision and abrasion fault feature extraction method based on fusion of vector spectrum homologous information and shaft vibration and shell vibration information

Abstract

The invention discloses a rotor collision and abrasion fault feature extraction method based on fusion of vector spectrum homologous information and shaft vibration and shell vibration information. The rotor collision and abrasion fault feature extraction method comprises the following steps that speed and displacement sensors are arranged at the same monitoring point at the same time to collect corresponding speed and displacement vibration signals; fusing the dual-channel speed signal and the displacement signal by using a vector spectrum technology; and according to the fused speed main vibration vector characteristics and displacement main vibration vector characteristics, establishing a rotor collision and abrasion fault diagnosis rule. According to the invention, the diagnosis accuracy of the rotor collision and abrasion fault is improved.

Description

Technical Field

[0001] This invention relates to the field of rotor rubbing fault diagnosis technology, and more specifically, to a method for extracting rotor rubbing fault features by fusing shaft vibration and shell vibration information with full vector spectrum homogeneous information. Background Technology

[0002] Wear-off faults are a common type of fault in large rotating machinery supported by sliding bearings (such as compressors and steam turbines). When extracting features from large bearing-supported coking equipment such as steam turbines and compressors, the information from two channels at the same cross-section often presents different fault characteristics, leading to misjudgments. In addition, due to limitations in sensor installation space, engineering applications often only monitor shaft vibration (displacement signal) for large rotating machinery supported by sliding bearings. However, wear-off faults are often accompanied by high-order harmonic characteristics of the power frequency. The shaft vibration information collected by the displacement sensor is only sensitive to low-frequency vibration components and can effectively capture low-frequency components of the power frequency fractional harmonics, but it is not sensitive to high-order harmonic characteristics and cannot reflect the high-order harmonic characteristics of wear-off faults. Summary of the Invention

[0003] To address the problems existing in the prior art, the purpose of this invention is to provide a method for extracting rotor rubbing fault features by fusing shaft vibration and shell vibration information with full vector spectrum homogeneous information, thereby improving the diagnostic accuracy of rotor rubbing faults.

[0004] The present invention adopts the following technical solution:

[0005] A method for extracting rotor rubbing fault features by fusing full vector spectrum homogeneous information with shaft vibration and shell vibration information includes the following steps:

[0006] S1. Simultaneously deploy velocity and displacement sensors at the same monitoring point to collect corresponding velocity and displacement vibration signals;

[0007] S2. The dual-channel velocity signal and displacement signal are fused using full vector spectrum technology.

[0008] S3. Establish rotor rubbing fault diagnosis rules based on the combined characteristics of velocity main vibration vector and displacement main vibration vector.

[0009] In one embodiment of the present invention, a rotor rubbing fault diagnosis rule is established based on the fused velocity principal vibration characteristics and displacement principal vibration characteristics.

[0010] The rotor rubbing fault diagnosis rule is as follows:

[0011] A rotor rubbing failure is determined to occur when all three of the following conditions are met simultaneously:

[0012] Condition 1: The amplitude of the passband of the displacement principal vibration vector spectrum is greater than or equal to the preset alarm value, or the amplitude of the passband of the velocity principal vibration vector spectrum is greater than or equal to the preset alarm value;

[0013] Condition 2: The ratio of the first harmonic amplitude to the passband amplitude in the principal displacement vector spectrum is ≥0.8;

[0014] Condition 3: The ratio of the sum of the amplitudes of the second, third, and fourth harmonics in the velocity master oscillation vector spectrum to the amplitude of the passband is ≥0.3.

[0015] Beneficial effects

[0016] This invention solves the problem of inconsistent information between two channels: by using full vector spectrum technology to fuse the dual-channel shaft vibration / shell vibration information of the same monitoring point, the one-sidedness of single-channel information is eliminated, and misjudgment caused by the difference in characteristics between the two channels is avoided;

[0017] Achieve comprehensive fault feature extraction: By combining displacement sensors (capturing low-frequency 1X features) and velocity sensors (capturing mid-to-high-frequency 2X-4X features), the complementary features of wear and tear faults are covered across the entire frequency band, overcoming the limitation of single shaft vibration monitoring being insensitive to high-order harmonics.

[0018] Improve diagnostic accuracy: Based on experimentally validated quantitative diagnostic rules (1X / pass frequency ≥ 0.8, (2X+3X+4X) / pass frequency ≥ 0.3), fault diagnosis becomes more objective and operable. Through simulation experiments, the diagnostic accuracy is improved.

[0019] This invention effectively solves the problems of inconsistent characteristics of dual-channel information on the same rotor interface and the difficulty of reflecting the high-order harmonic characteristics of rotor rubbing faults with single shaft vibration information, and provides an innovative method to improve the diagnostic accuracy of rotor rubbing faults. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the installation of a dual-channel sensor according to an embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of rotor node motion according to an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of the rotor center motion trajectory according to an embodiment of the present invention;

[0023] Figure 4(a) is a spectrum diagram of dual-channel shaft vibration information in the X direction collected from the same cross section of the rotor according to an embodiment of the present invention;

[0024] Figure 4(b) is a spectrum diagram of dual-channel shaft vibration information in the Y direction collected from the same cross section of the rotor according to an embodiment of the present invention;

[0025] Figure 4(c) is a schematic diagram of the main vibration vector of the rotor after dual-channel full-vector fusion according to an embodiment of the present invention;

[0026] Figure 4(d) is a schematic diagram of the secondary vibration vector of the rotor after dual-channel full-vector fusion in an embodiment of the present invention;

[0027] Figure 5(a) is a spectrum diagram of dual-channel shell vibration information in the X direction collected from the same cross section of the rotor according to an embodiment of the present invention;

[0028] Figure 5(b) is a spectrum diagram of dual-channel shell vibration information in the Y direction collected from the same cross section of the rotor according to an embodiment of the present invention;

[0029] Figure 5(c) is a schematic diagram of the main vibration vector of the shell vibration information after dual-channel full-vector fusion of the rotor according to an embodiment of the present invention;

[0030] Figure 5(d) is a schematic diagram of the shell vibration information of the rotor after dual-channel full-vector fusion according to an embodiment of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0032] As shown in the figure, this invention discloses a method for extracting rotor rubbing fault features by fusing shaft vibration and shell vibration information with full vector spectrum homogeneous information, including the following steps:

[0033] S1. Simultaneously deploy velocity and displacement sensors at the same monitoring point to collect corresponding velocity and displacement vibration signals;

[0034] S2. The dual-channel velocity signal and displacement signal are fused using full vector spectrum technology.

[0035] S3. Establish rotor rubbing fault diagnosis rules based on the combined characteristics of velocity main vibration vector and displacement main vibration vector.

[0036] In step S1, the sensors for acquiring velocity and displacement vibration signals are arranged as follows:

[0037] Two speed sensors and two displacement sensors are simultaneously arranged at the bearing support of the rotor.

[0038] Two speed sensors are installed at 90 degrees to each other. The X-direction channel is installed before the Y-direction channel according to the rotor rotation direction to collect dual-channel co-source shell vibration information. Two displacement sensors are also installed at 90 degrees to collect dual-channel co-source shaft vibration information, in the same order as the speed sensors.

[0039] In step S2, the dual-channel velocity signal and displacement signal are fused using the full vector spectrum technique to obtain the principal velocity vector and the principal displacement vector, respectively. The specific theoretical calculation of the dual-channel full vector spectrum is as follows:

[0040] The equation of motion for a rotor undergoing steady-state vortexing with an angular velocity Ω in two-dimensional space is:

[0041] (1)

[0042] in:

[0043] Equation (1) eliminates time t, yielding the trajectory equation (ellipse equation) of the disk center:

[0044] (2)

[0045] The two-dimensional total vector spectrum characteristic parameters Ra, Rb, a, and Φ are obtained as follows:

[0046] (3)

[0047] (4)

[0048] (5) (6)

[0049] Where: R a —Main oscillator, R a The semi-major axis of the ellipse

[0050] R b —Paramisonic vector, R b The minor semi-axis of the ellipse

[0051] a—Sagittal angle, where a is the angle between the semi-major axis Ra and the x-axis.

[0052] Φ—Sagittal phase, where Φ is the phase angle of the disk center moving along an elliptical trajectory.

[0053] The method for obtaining the eigenvalues ​​of the two-dimensional full vector spectrum is as follows:

[0054] By R a R b From the formulas for calculating a and Φ, we know that we need to obtain x. c x s y c y s value

[0055] From equation (1), we can obtain:

[0056] (7)

[0057] Performing a Discrete Fourier Transform (DFT) on the above equation yields:

[0058] (8)

[0059] Similarly, we can conclude that:

[0060] (9)

[0061] The obtained x c x s y c y s Substitute R a R b These characteristic quantities are obtained by using the formulas for calculating a and Φ.

[0062] Using a certain type of steam turbine rotor as the test object, sensors were arranged at its bearing support (same monitoring section):

[0063] Speed ​​sensor: Two piezoelectric speed sensors are used, installed at a 90° angle, with the X-axis leading the Y-axis by 15mm along the rotor rotation direction;

[0064] Displacement sensor: Two eddy current displacement sensors are used, installed at a 90° angle and arranged in the same cross section as the velocity sensor.

[0065] Then signal acquisition and fusion are performed.

[0066] Signal acquisition: The shaft vibration (displacement) and shell vibration (velocity) signals in the X and Y directions are synchronously acquired through a data acquisition card (sampling frequency 1024Hz) for 10 seconds.

[0067] Full vector spectrum fusion: Perform DFT on the acquired dual-channel displacement signal to obtain... , , Substituting into equation (3), the principal displacement vector is calculated. Its 1X amplitude is 0.048A, the passband amplitude is 0.058A, and 1X / passband ≈ 0.827 (satisfying ≥ 0.8). A DFT is performed on the dual-channel velocity signal to obtain... , , The velocity principal oscillation is calculated, with an amplitude of 0.62 mm / s for 2X+3X+4X and an amplitude of 1.98 mm / s for the passband. The ratio is approximately 0.313 (satisfying ≥0.3).

[0068] Fault diagnosis results

[0069] Basic conditions: Displacement principal vibration amplitude at the pass frequency ≥ alarm value 0.058A, velocity principal vibration amplitude at the pass frequency ≥ alarm value 1.98mm / s, satisfying the "OR" condition;

[0070] Low-frequency condition: 1X / passband ≈ 0.827 ≥ 0.8, satisfied;

[0071] The condition for higher harmonics is: (2X+3X+4X) / passband ≈ 0.313 ≥ 0.3, which is satisfied.

[0072] This invention establishes rotor rubbing fault diagnosis rules based on the fused principal vibration characteristics of velocity and displacement. The rotor rubbing fault diagnosis rules are as follows:

[0073] A rotor rubbing failure is determined to occur when all three of the following conditions are met simultaneously:

[0074] Condition 1: The amplitude of the passband of the displacement principal vibration vector spectrum is greater than or equal to the preset alarm value, or the amplitude of the passband of the velocity principal vibration vector spectrum is greater than or equal to the preset alarm value;

[0075] Condition 2: The ratio of the 1x frequency amplitude to the passband amplitude in the displacement principal oscillation vector spectrum is ≥0.8; Figure 4(c) effectively reflects the low-frequency characteristics of 1x during a rubbing fault. Simultaneously, Figure 4(c) effectively reflects the high-frequency characteristics of higher harmonics such as 2x, 3x, and 4x during a rubbing fault. Combining Figure 4(c) and Figure 5(c), it can be determined that the rotor has experienced a dynamic-static rubbing fault.

[0076] Condition 3: The ratio of the sum of the amplitudes of the 2nd (2x) frequency, the 3rd (3x) frequency, and the 4th (4x) frequency in the velocity master oscillation vector spectrum to the passband amplitude is ≥0.3.

[0077] The above are quantitative indicators, as follows:

[0078] Principal vibration vector of shaft vibration: 1X / pass frequency amplitude >= 0.8

[0079] Velocity master oscillation vector spectrum: (2X amplitude + 3X amplitude + 4X amplitude) / passband amplitude >= 0.3.

[0080] The rotor rubbing fault diagnosis rules are converted into an executable program through pseudocode, which is adapted to the development language of the fault diagnosis system.

[0081] The above description is merely a preferred embodiment of the present invention; however, the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and its improved concepts, should be covered within the scope of protection of the present invention.

Claims

1. A method for extracting rotor rubbing fault features by fusing shaft vibration and shell vibration information with full vector spectrum homogeneous information, characterized in that: Includes the following steps: S1. Simultaneously deploy velocity and displacement sensors at the same monitoring point to collect corresponding velocity and displacement vibration signals; S2. The dual-channel velocity signal and displacement signal are fused using full vector spectrum technology respectively; S3. Establish rotor rubbing fault diagnosis rules based on the combined characteristics of velocity main vibration vector and displacement main vibration vector.

2. The rotor rubbing fault feature extraction method according to claim 1, which integrates full vector spectrum homogeneous information fusion of shaft vibration and shell vibration information, is characterized in that: In step S1, the sensors for acquiring velocity and displacement vibration signals are arranged as follows: Two speed sensors and two displacement sensors are simultaneously arranged at the bearing support of the rotor.

3. The rotor rubbing fault feature extraction method according to claim 2, which integrates full vector spectrum homogeneous information and shaft vibration and shell vibration information, is characterized in that: In step S2, the dual-channel velocity signal and displacement signal are fused using the full vector spectrum technique to obtain the principal velocity vector and the principal displacement vector, respectively. The theoretical calculation of the dual-channel full vector spectrum is as follows: The equation of motion for a rotor undergoing steady-state vortexing with an angular velocity Ω in two-dimensional space is: (1) in: , Equation (1) eliminates time t, yielding the trajectory equation (ellipse equation) of the disk center: (2) The two-dimensional total vector spectrum characteristic parameters Ra, Rb, a, and Φ are obtained as follows: (3) (4) (5) (6) Where: R a —Main oscillator, R a The semi-major axis of the ellipse R b —Paramisonic vector, R b The minor semi-axis of the ellipse a—Sagittal angle, where a is the angle between the semi-major axis Ra and the x-axis. Φ—Sagittal phase, where Φ is the phase angle of the disk center moving along an elliptical trajectory.

4. The rotor rubbing fault feature extraction method according to claim 3, which fuses shaft vibration and shell vibration information with full vector spectrum homogeneous information, is characterized in that: The method for obtaining the eigenvalues ​​of the two-dimensional full vector spectrum is as follows: By R a R b From the formulas for calculating a and Φ, we know that we need to obtain x. c x s y c y s value From equation (1), we can obtain: (7) Performing a Discrete Fourier Transform (DFT) on the above equation yields: (8) Similarly, we can conclude that: (9) The obtained x c x s y c y s Substitute R a R b These characteristic quantities are obtained by using the formulas for calculating a and Φ.

5. The rotor rubbing fault feature extraction method according to claim 4, which fuses shaft vibration and shell vibration information with full vector spectrum homogeneous information, is characterized in that: Based on the combined characteristics of the velocity principal vibration vector and the displacement principal vibration vector, a rotor rubbing fault diagnosis rule is established.

6. The rotor rubbing fault feature extraction method according to claim 5, which fuses shaft vibration and shell vibration information with full vector spectrum homogeneous information, is characterized in that: The rotor rubbing fault diagnosis rule is as follows: A rotor rubbing failure is determined to occur when all three of the following conditions are met simultaneously: Condition 1: The amplitude of the passband of the displacement principal vibration vector spectrum is greater than or equal to the preset alarm value or the amplitude of the passband of the velocity principal vibration vector spectrum is greater than or equal to the preset alarm value; Condition 2: The ratio of the first harmonic amplitude to the passband amplitude in the principal displacement vector spectrum is ≥0.8; Condition 3: The ratio of the sum of the amplitudes of the second, third, and fourth harmonics in the velocity master oscillation vector spectrum to the amplitude of the passband is ≥0.

3.

7. The rotor rubbing fault feature extraction method according to claim 6, which fuses shaft vibration and shell vibration information with full vector spectrum homogeneous information, is characterized in that: The rotor rubbing fault diagnosis rules are converted into an executable program through pseudocode, which is adapted to the development language of the fault diagnosis system.

8. The rotor rubbing fault feature extraction method according to claim 2, which integrates full vector spectrum homogeneous information and shaft vibration and shell vibration information, is characterized in that: Two speed sensors are installed at 90 degrees. The X-direction channel is installed before the Y-direction channel according to the rotor rotation direction to collect dual-channel co-source shell vibration information. Two displacement sensors are also installed at 90 degrees to collect dual-channel co-source shaft vibration information, in the same order as the speed sensors.

Images