A bearing running state fault size diagnosis method

Abstract

The application relates to a bearing operation state fault size diagnosis method and relates to the field of bearing fault detection. The application can solve the problem that existing bearing diagnosis methods cannot distinguish bearing fault sizes, thereby failing to refine the bearing fault development state and further causing low bearing fault diagnosis accuracy. The application comprises the following steps: obtaining a bearing weak magnetic signal, obtaining the running speed of bearing rolling bodies by using the bearing weak magnetic signal; performing self-adaptive decomposition on the bearing weak magnetic signal to obtain an intrinsic mode component, and obtaining an effective bearing vibration mode component by using the intrinsic mode component; obtaining waveform features of the bearing rolling bodies in a fault region and a non-fault region by using the effective bearing vibration mode component; obtaining the time when the bearing rolling bodies enter and leave the fault region by using the waveform features, thereby obtaining the running time of the bearing rolling bodies in the fault region; and obtaining the bearing fault size by using the running time of the bearing rolling bodies in the fault region and the running speed of the bearing rolling bodies. The application is used for obtaining the bearing operation state fault size.

Description

Technical Field

[0001] This invention relates to the field of bearing fault detection, and in particular to a method for diagnosing the dimensional faults in bearing operation. Background Technology

[0002] Rolling bearings are crucial components of rotating machinery, directly impacting its safe and stable operation. Their applications span numerous fields, including energy, metallurgy, mining, aviation, and aerospace. For aircraft engine bearings, in particular, the harsh environments of high temperature, high speed, heavy load, and oil mist pose significant risks. Failures in aircraft bearings can easily lead to serious flight accidents, ranging from increased engine vibration and rotor wear to severe engine damage, such as seizure due to main bearing wear. Therefore, bearing fault diagnosis is of paramount importance.

[0003] Currently, bearing fault diagnosis primarily focuses on analyzing relevant signals collected through vibration monitoring, acoustic emission monitoring, and oil monitoring to assess the operating status of rolling bearings. Vibration monitoring methods utilize time-domain diagnostic parameters such as RMS, RMS, peak value, kurtosis, peak factor, waveform factor, and margin indices; frequency-domain monitoring features include power spectrum, amplitude spectrum, refined spectrum, and cepstrum. Acoustic emission is often used to diagnose early-stage bearing faults, with research methods primarily focusing on time-varying kurtosis and wavelet packet decomposition. Oil monitoring includes spectral analysis, ferrography, and abrasive particle detection. While these methods can diagnose bearing faults from different perspectives and establish a quantitative relationship between bearing operating status and signal analysis, this quantitative assessment is limited to a rough evaluation of the bearing's operating state, such as the initial fault, the fault development stage, and the complete failure state. It cannot determine the size of the bearing fault, leading to a lack of refinement in the fault development stage and consequently, low accuracy in bearing fault diagnosis. Summary of the Invention

[0004] The purpose of this invention is to solve the problem that existing bearing diagnostic methods cannot distinguish the size of bearing faults, thus failing to refine the development state of bearing faults and resulting in low accuracy of bearing fault diagnosis. Therefore, this invention proposes a method for diagnosing bearing fault size in operating conditions.

[0005] The specific process of a method for diagnosing bearing operating condition fault dimensions is as follows:

[0006] Step 1: Obtain the bearing magnetic weakening signal, use the bearing magnetic weakening signal to obtain the characteristic frequency of the bearing rolling element operation, and then use the characteristic frequency of the bearing rolling element operation to obtain the bearing rolling element speed.

[0007] Step 2: Adaptively decompose the bearing field weakening signal obtained in Step 1 to obtain the intrinsic mode components, and use the intrinsic mode components to obtain the effective bearing vibration mode components.

[0008] Step 3: Perform Hilbert transform on the effective bearing vibration mode components obtained in Step 2 to obtain the waveform characteristics of the bearing rolling elements in the fault region and non-fault region. Then, use the obtained waveform characteristics to obtain the time when the bearing rolling elements enter and leave the fault region, thereby obtaining the time when the bearing rolling elements run in the fault region.

[0009] Step 4: Use the running time of the bearing rolling elements in the fault area obtained in Step 3 and the running speed of the bearing rolling elements obtained in Step 1 to obtain the bearing fault size.

[0010] Preferably, obtaining the bearing magnetic weakening signal in step one specifically involves:

[0011] First, install a weak magnetic field detection sensor:

[0012] A hole of the same size as the sensor is drilled in the bearing housing along the radial direction of the bearing, and the weak magnetic field detection sensor is placed in the hole;

[0013] Then, the weak magnetic field detection sensor is made to be non-contact with the outer ring of the bearing, and the weak magnetic field signal of the bearing is obtained by the weak magnetic field detection sensor.

[0014] Preferably, the step one of obtaining the characteristic frequency of the bearing rolling element operation by using the bearing magnetic weakening signal specifically involves: performing a Fourier transform on the bearing magnetic weakening signal to obtain the characteristic frequency of the bearing rolling element operation.

[0015] Furthermore, the step one of obtaining the bearing rolling element speed using the characteristic frequency of the bearing rolling element's operation specifically involves:

[0016] v = f c ×60

[0017] Among them, f c ν is the characteristic frequency of the bearing rolling element, and v is the speed of the bearing rolling element.

[0018] Furthermore, the step one of obtaining the bearing rolling element speed using the characteristic frequency of the bearing rolling element's operation specifically involves:

[0019] v = f c ×60

[0020] Among them, f c ν is the characteristic frequency of the bearing rolling element, and v is the speed of the bearing rolling element.

[0021] Furthermore, in step two, the bearing field weakening signal obtained in step one is adaptively decomposed to obtain intrinsic mode components, and the effective bearing vibration mode components are obtained using the intrinsic mode components. Specifically, this involves:

[0022] Step 2: The bearing field weakening signal obtained in Step 1 is adaptively decomposed using an adaptive empirical mode decomposition method to obtain the intrinsic mode components IMF1, IMF2, ..., IMF1. n ;

[0023] Where n is the total number of intrinsic mode components;

[0024] Step 22: Obtain the effective bearing vibration mode components from the intrinsic mode components based on the energy maximum method:

[0025] Calculate the mean energy of each intrinsic mode component obtained in step two, and compare the mean energies of all intrinsic mode components. The intrinsic mode component corresponding to the maximum mean energy is the effective bearing vibration mode component.

[0026] Furthermore, the mean energy of each intrinsic mode component is obtained by the following formula:

[0027]

[0028] Where M is the length of the bearing field weakening signal, x i Let i be the signal point corresponding to each intrinsic mode component in the bearing field weakening signal, and i is the sequence number of the signal point of each intrinsic mode component in the bearing field weakening signal.

[0029] Furthermore, the step four, which involves using the running time of the bearing rolling elements in the fault region obtained in step three and the running speed of the bearing rolling elements obtained in step one to obtain the bearing fault size, specifically involves:

[0030]

[0031] Among them, t f Let v be the running time of the bearing rolling element in the fault region, v be the running speed of the bearing rolling element, R be the inner ring radius of the bearing, and r be the radius of the bearing rolling element.

[0032] The beneficial effects of this invention are as follows:

[0033] This invention utilizes effective bearing vibration mode components to obtain the time when the bearing rolling elements enter and leave the fault region, thereby obtaining the running time of the rolling elements in the fault region. Then, this invention combines the running time of the bearing rolling elements in the fault region with the extracted rolling element speed to obtain the fault size of the bearing's operating state. By discriminating the bearing fault size, this invention further refines the development state of the bearing fault, improving the accuracy of bearing fault diagnosis. Attached Figure Description

[0034] Figure 1 A flowchart of the bearing operating condition fault size diagnosis method provided by the present invention;

[0035] Figure 2 This is a time-domain diagram of the monitoring signal in an embodiment of the present invention;

[0036] Figure 3 This is the Fourier transform of the weak magnetic field detection signal of the present invention;

[0037] Figure 4 This is a diagram showing the selection of effective vibration mode components in this invention;

[0038] Figure 5 The Hilbert transform of the effective vibration mode components of this invention;

[0039] Figure 6 This is a diagram of the bearing fault size diagnosis model of the present invention;

[0040] Figure 7 This is the bearing fault size diagnosis result of the present invention. Detailed Implementation

[0041] Specific implementation method one: as follows Figure 1 As shown, the specific process of a bearing operating condition fault size diagnosis method in this embodiment is as follows:

[0042] Step 1: Obtain the bearing's weak magnetic field signal using a non-contact magnetic field detection sensor. Perform a Fourier transform on the bearing's weak magnetic field signal to obtain the characteristic frequency of the bearing's rolling element operation. Then, use the characteristic frequency of the bearing's rolling element operation to obtain the bearing's rolling element speed. Specifically:

[0043] Step 11: Obtain the bearing's weak magnetic signal using a weak magnetic field detection sensor for non-contact detection.

[0044] First, install a weak magnetic field detection sensor:

[0045] A hole identical to the sensor is drilled along the radial direction of the bearing on the bearing housing of the fixed bearing, and the weak magnetic field detection sensor is placed in the hole;

[0046] Then, the weak magnetic field detection sensor is made to be non-contact with the outer ring of the bearing, and the weak magnetic field detection sensor is used to pick up the characteristic information of the bearing's disturbance of the geomagnetic field, and obtain the weak magnetic signal of the bearing.

[0047] Steps 1 and 2: Perform a Fourier transform on the bearing magnetic weakening signal to obtain the characteristic frequencies of the bearing rolling elements. Then, use the characteristic frequencies of the bearing rolling elements to obtain the bearing rolling element speed, as shown in the following formula:

[0048] v = fc ×60

[0049] Among them, f c ν is the characteristic frequency of the bearing rolling element, and v is the speed of the bearing rolling element.

[0050] Step 2: Adaptive decomposition of the bearing field weakening signal obtained in Step 1 is performed using the adaptive empirical mode decomposition method (CEEMDAN) to obtain the intrinsic mode components. Then, based on the energy maximum rule, the effective bearing vibration mode components are obtained using the intrinsic mode components.

[0051] Step 2: 1. Decompose the bearing field weakening signal obtained in Step 1 using CEEMDAN to obtain the intrinsic mode components IMF1, IMF2, ..., IMF1. n ;

[0052] n is the total number of intrinsic mode components;

[0053] Step 22: Calculate the mean energy of each intrinsic mode component obtained in Step 21, and compare the mean energies of all intrinsic mode components. The intrinsic mode component corresponding to the maximum mean energy is the effective bearing vibration mode component.

[0054] The average energy is calculated using the following formula:

[0055]

[0056] Where M is the length of the bearing field weakening signal, x i Let i be the signal point corresponding to each intrinsic mode component in the bearing field weakening signal, and i is the sequence number of the signal point of each intrinsic mode component in the bearing field weakening signal.

[0057] Step 3: Perform Hilbert transform on the effective bearing vibration mode components obtained in Step 2 to obtain the waveform characteristics of the bearing rolling elements in the fault region and non-fault region. Use the obtained waveform characteristics to obtain the time when the bearing rolling elements enter and leave the fault region, thereby obtaining the time when the bearing rolling elements run in the fault region.

[0058] Step 4: Using the running time of the bearing rolling elements in the fault area obtained in Step 3 and the running speed of the bearing rolling elements obtained in Step 1, obtain the bearing fault size:

[0059]

[0060] Among them, t f Let v be the running time of the bearing rolling element in the fault region, v be the running speed of the bearing rolling element, R be the inner ring radius of the bearing, and r be the radius of the bearing rolling element.

[0061] Example: To verify the technical effect of the present invention, the following experiments were conducted:

[0062] Step S1: Use a weak magnetic field detection sensor to non-contactly detect the speed of the bearing rolling elements and the characteristics of bearing vibration information;

[0063] A pre-set 1.705mm fault defect was used in the bearing. A weak magnetic field sensor was employed on the bearing test bench to detect the bearing's disturbance to the Earth's magnetic field during operation. The data acquisition card's data acquisition frequency was 25.6Hz. The bearing was set to accelerate first and then decelerate at speeds of 1010 rpm (fr1), 1610 rpm (fr2), 1810 rpm (fr3), 2710 rpm (fr4), 1810 rpm (fr3), 1610 rpm (fr2), and 1010 rpm (fr1), with each speed stage lasting 60 seconds. This ensured that the bearing's operation generated a weak magnetic field signal indicating its disturbance to the Earth's magnetic field. Figure 2 As shown. Due to the large amount of data, this embodiment selects the signal from one speed change phase for analysis. From Figure 2 It can be observed that the amplitude of the weak magnetic signal decreases as the rotational speed increases.

[0064] Step S2, the bearing rolling element speed extraction stage, performs Fourier transform on the acquired weak magnetic signal to extract the characteristic frequency of the rolling element operation and obtain the rolling element speed;

[0065] Perform a Fourier transform on the detected weak magnetic signal, such as Figure 3 As shown in the figure, the Fourier transform of the weak magnetic field detection signal contains multiple information components: the rotational frequency (fr), second harmonic (2fr), third harmonic (3fr), and sixth harmonic (6fr) of the bearing inner ring, the multiple rotational frequencies (fm) of the rolling elements, and the corresponding rotational frequency fc of the rolling elements is fm / N, where N is the number of rolling elements.

[0066] Step S3: During the rolling element running time extraction stage in the fault area, the detected weak magnetic signal is adaptively decomposed using the adaptive empirical mode decomposition method (CEEMDAN), and the effective bearing vibration mode components are obtained based on the maximum energy rule.

[0067] The CEEMDAN algorithm is used to adaptively decompose the weak magnetic field detection signal, calculate the mean energy of each eigenmode component, and select the eigenmode component with the maximum energy as the effective vibration mode component. The flowchart is shown below. Figure 4 As shown, the method for calculating the maximum mean energy is as follows:

[0068]

[0069] Where M is the number of data points analyzed, x i This is a signal point.

[0070] Step S4: Perform Hilbert transform on the extracted effective mode components to obtain the waveform characteristics of the bearing rolling elements in the fault area and non-fault area. For the waveform after Hilbert transform, identify the time feature points of the bearing rolling elements entering and leaving the fault area.

[0071] The effective vibration mode components are subjected to Hilbert transform, and the transformed graph is shown below. Figure 5 As shown in the figure, the Hilbert transform graphs of the rolling element in the bearing fault region and the non-fault region are significantly different. By identifying the time points in the fault region and the non-fault region, the running time information of the rolling element in the fault region (t) can be obtained. f ).

[0072] Step S5: Establish a bearing fault size model based on the bearing's geometric and operational characteristics, using the rolling element velocity (v) and time (t) obtained in steps S2 and S4. f ), to diagnose bearing fault dimensions.

[0073] Combining bearing geometric features (such as) Figure 6 As shown in the figure, a diagnostic model for bearing fault dimensions is established, as shown in formula (2):

[0074]

[0075] from Figure 6 It can be observed that the fault size diagnosis of the bearing is related to the running speed of the rolling elements, the running time of the rolling elements in the fault area, the inner diameter of the bearing inner ring, and the radius of the rolling elements. The rolling element speed (v) and the running time (t) in the fault area obtained in steps S2 and S4 are combined. f By inputting the bearing diagnostic model from S5, the bearing fault dimensions can be obtained, such as... Figure 7 As shown in the figure, the diagnosed bearing fault size deviates from the actual preset size. The main reason is that the running state of the bearing rolling elements in the fault area is inconsistent, resulting in different estimates of the running time in the fault area. The maximum estimation error is within 11%. Among them, fr1, fr2, fr3, and fr4 correspond to the velocity spectrum set in step 1 of this embodiment. The bearing fault size estimation within the error range has certain guiding significance for the actual bearing fault diagnosis.

Claims

1. A bearing running state failure size diagnosing method characterized by The method specifically comprises the following steps: Step one, obtaining a bearing weak magnetic signal, using the bearing weak magnetic signal to obtain a characteristic frequency of bearing rolling body operation, and then using the characteristic frequency of bearing rolling body operation to obtain a bearing rolling body operation speed; Step two, performing adaptive decomposition on the bearing weak magnetic signal obtained in step one to obtain an intrinsic mode component, and using the intrinsic mode component to obtain an effective bearing vibration mode component; Step three, performing Hilbert transform on the effective bearing vibration mode component obtained in step two to obtain waveform features of the bearing rolling body in a fault region and a non-fault region, and then using the obtained waveform features to obtain time points at which the bearing rolling body enters the fault region and leaves the fault region, so as to obtain a running time of the bearing rolling body in the fault region; Step four, using the running time of the bearing rolling body in the fault region obtained in step three and the bearing rolling body operation speed obtained in step one to obtain a bearing fault size.

2. The method of claim 1, wherein: In the step one, the bearing weak magnetic signal is obtained in the following manner: First, installing a weak magnetic detection sensor: A hole with the same size as the sensor is formed on a bearing seat of a fixed bearing along a radial direction of the bearing, and the weak magnetic detection sensor is placed in the hole; Then, the weak magnetic detection sensor is placed in non-contact with an outer ring of the bearing, and the bearing weak magnetic signal is obtained by using the weak magnetic detection sensor.

3. A method of diagnosing the size of a fault in the operation of a bearing according to claim 2 or 1, characterized in that: In the step one, the characteristic frequency of the bearing rolling body operation is obtained by performing Fourier transform on the bearing weak magnetic signal.

4. The method of claim 3, wherein: In the step one, the bearing rolling body operation speed is obtained by using the characteristic frequency of the bearing rolling body operation. v = f c x 60 where f c is the characteristic frequency of the bearing rolling element operation, v is the speed of the bearing rolling element operation.

5. The bearing operational state failure size diagnostic method according to claim 4, characterized by: In the step two, the effective bearing vibration mode component is obtained by performing adaptive decomposition on the bearing weak magnetic signal obtained in step one, and the effective bearing vibration mode component is obtained by using the intrinsic mode component, in the following manner: Step two, the bearing weak magnetic signal obtained in step one is adaptively decomposed by using the adaptive empirical mode decomposition method to obtain intrinsic mode components IMF1, IMF2, …, IMF n ; Wherein, n is the total number of the intrinsic mode components; Step two one, obtaining the effective bearing vibration mode component in the intrinsic mode component based on an energy maximum value method: The mean energy of each intrinsic mode component obtained in step two one is calculated, and the mean energies of all the intrinsic mode components are compared, and the intrinsic mode component corresponding to the maximum mean energy is the effective bearing vibration mode component.

6. A method of diagnosing a fault size of a bearing operating state according to claim 5, characterized by: The mean energy of each intrinsic mode component is obtained by the following formula: Wherein, M is the length of bearing flux-weakening signal, x i is the signal point corresponding to each eigenmode component in the bearing flux-weakening signal, and i is the serial number of the signal point of each eigenmode component in the bearing flux-weakening signal.

7. A method of diagnosing a fault size of a bearing operating state according to claim 6, characterized by: In the step four, the bearing fault size is obtained by using the running time of the bearing rolling body in the fault region obtained in step three and the bearing rolling body operation speed obtained in step one, in the following manner: wherein t f is the running time of the bearing rolling element in the fault zone, v is the speed at which the bearing rolling element is running, R is the radius of the inner ring of the bearing, and r is the radius of the bearing rolling element.

Images