Bearing fault detection method based on adaptive continuous sampling

Abstract

The invention discloses a bearing fault detection method based on self-adaptive continuous sampling, and the method comprises the steps: selecting the type of a sensor, and determining the installation position and detection parameters of the sensor; synchronously acquired vibration data are acquired; reconstructing synchronously acquired vibration data; calculating fault features; and judging whether a fault occurs or not and adaptively adjusting the synchronously acquired data volume of each rotation of the rotating shaft in next fault detection. According to the method, under a certain reliability requirement, the data acquisition amount in bearing fault detection can be reduced, and the data transmission, processing and storage costs are reduced.

Description

Technical Field

[0001] The invention belongs to the field of bearing fault detection and diagnosis, and in particular relates to a bearing fault detection method based on adaptive continuous sampling. Background Art

[0002] In modern industrial production, the efficient and stable operation of mechanical equipment is one of the key factors to ensure production efficiency and product quality. As the core component of the mechanical system, bearings bear important functions such as supporting rotating shafts, reducing friction, and transmitting loads. However, since bearings are under high load and high-speed operation for a long time, they are prone to various failures. These failures will not only seriously affect the performance and life of the equipment, but may also cause major safety accidents. Therefore, timely and accurate bearing fault detection and early detection of fault conditions are of great significance to achieving bearing health management and reducing maintenance costs.

[0003] Traditional bearing fault detection methods mainly rely on manual regular inspections and empirical judgments. This method is highly subjective, inefficient, and difficult to capture early signs of faults. With the advancement of science and technology, especially the rapid development of signal processing technology, pattern recognition technology, and artificial intelligence technology, a series of advanced bearing fault detection technologies have emerged, greatly improving the accuracy and real-time performance of fault detection.

[0004] In the field of bearing fault detection, the fault detection method based on vibration signal analysis is one of the most common methods. Through the sensors installed on the bearings or mechanical equipment, the vibration signals of the bearings during operation are collected, and then the fault features are extracted from the complex signals by means of time domain analysis, frequency domain analysis, time-frequency analysis, etc.

[0005] Traditional bearing fault detection methods often use a continuous sampling method with a fixed sampling frequency, which has the disadvantages of redundant sampling data, long data transmission delay, and low transmission rate. This sampling method is no longer suitable for bearing fault diagnosis. Studies have found that bearing degradation has certain gradual characteristics. Therefore, researchers propose to use a periodic detection method to detect bearing faults. Although this detection method can significantly reduce data redundancy, sampling costs, transmission costs, etc., it cannot guarantee the reliability of fault detection. Bearing faults may occur outside the period of regular detection.

[0006] In order to overcome the shortcomings of traditional continuous sampling detection and periodic detection in bearing fault detection, the present invention patent discloses a bearing fault detection method based on adaptive continuous sampling. Summary of the invention

[0007] Purpose of the invention: The present invention proposes a bearing fault detection method based on adaptive continuous sampling, which can significantly reduce the amount of data collection and reduce the cost of data transmission, processing and storage while ensuring the reliability of fault detection.

[0008] Technical solution: The present invention proposes a bearing fault detection method based on adaptive continuous sampling, which specifically includes the following steps:

[0009] Step (1): Select the sensor type, determine the sensor installation location and detection parameters;

[0010] Step (2): Acquire synchronously collected vibration data;

[0011] Step (3): reconstructing synchronously collected vibration data;

[0012] Step (4): Calculate fault characteristics;

[0013] Step (5): Determine whether a fault has occurred and adaptively adjust the amount of synchronously collected data per rotation of the shaft for the next fault detection.

[0014] Furthermore, the step (1) of selecting the sensor type includes: the selected sensor is a vibration sensor and a rotation speed sensor.

[0015] Furthermore, the step (1) of determining the sensor installation position includes: the vibration sensor is installed near the measured bearing to measure the bearing vibration data; the speed sensor is installed near the shaft to measure the shaft speed signal.

[0016] Furthermore, the step (1) of determining the detection parameters includes: determining the initial amount N0, the minimum amount N1 and the maximum amount N2 of the synchronously collected data amount N per rotation of the shaft, the sampling revolution number k, and two thresholds T1 and T2; when the detected bearing is a brand new bearing, N0 is set to N min ; When the health status of the detected bearing is unknown, set Among them, round() is a rounding function; the value range of the sampling speed k is 5 to 20; the two thresholds T1 and T2 are determined according to the mean μ0 and standard deviation σ0 of the fault characteristics when the bearing is fault-free, T1 = μ0 and T2 = μ0 + 3σ0.

[0017] Further, the step (2) includes: based on the vibration sensor and the speed sensor, according to the synchronously collected data amount N per rotation of the shaft and the sampling rotation number k, obtaining the synchronously collected vibration data S of the bearing with a data length of kN, and the synchronously collected vibration signal S can be expressed as [x1, x2, x3, ..., x kN ].

[0018] Furthermore, the step (3) includes: reconstructing the one-dimensional synchronous vibration data S into N groups of one-dimensional synchronous vibration data Y = [S1, S2, S3, ..., S i ,…,S N ], where i = 1, 2, 3, ..., N, the i-th group of one-dimensional synchronous vibration data S i It can be expressed as [x i ,x (i+N) ,x (i+2N) ,...,x (i+(k-1)N) ].

[0019] Furthermore, the step (4) comprises: respectively calculating the standard deviation Z of N groups of one-dimensional synchronous vibration data Y = [Z1, Z2, Z3, ..., Z i ,…,Z N ], the standard deviation of the i-th group of one-dimensional synchronous vibration data is

[0020]

[0021] Among them, μ i is the mean of the i-th group of one-dimensional synchronous vibration data; then the mean μ of the standard deviation Z is calculated and used as the fault feature,

[0022]

[0023] Furthermore, the step (5) includes: judging whether the bearing fault occurs according to the obtained fault feature μ and the two set thresholds T1 and T2, and adaptively adjusting the amount of synchronously collected data N per rotation of the shaft for the next fault detection; when the fault feature μ < T1, it indicates that the bearing is in a healthy state, then the amount of synchronously collected data N per rotation of the shaft for the next fault detection is adjusted to the minimum value N1, and the process is turned to step (2) to continue the bearing fault detection process; when the fault feature T1 ≤ μ ≤ T2, it is impossible to reliably determine whether the bearing fault occurs, and at this time the amount of synchronously collected data N per rotation of the shaft for the next fault detection is adjusted to Among them, the function ceil() is a rounding up function, and goes to step (2) to continue the bearing fault detection process; when the fault feature μ>T2, it indicates that the bearing fault has occurred, and the bearing fault detection process is stopped.

[0024] Compared with the prior art, the method of the present invention has the following beneficial effects: (1) the method of the present invention can adaptively adjust the amount of data synchronously collected per rotation of the shaft according to the degradation state of the bearing being tested; (2) while ensuring the reliability of bearing testing, the amount of data collected is significantly reduced, and the data transmission cost, processing cost and storage cost are reduced; (3) the method of the present invention is not affected by changes in the shaft speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of the method of the present invention;

[0026] Figure 2 This is the time domain diagram of the fault-free simulation signal of the gearbox bearing;

[0027] Figure 3 Time domain diagram of the simulated signal of a gearbox bearing fault;

[0028] Figure 4 This is a schematic diagram of the test results when the gearbox bearing has no faults;

[0029] Figure 5 It is a schematic diagram of the number of synchronously sampled data per shaft revolution when the gearbox bearing is fault-free;

[0030] Figure 6 This is a schematic diagram of the detection results when the gearbox bearing fails;

[0031] Figure 7 This is a schematic diagram of the number of synchronously sampled data per shaft revolution when a gearbox bearing fails. DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings.

[0033] The present invention discloses a bearing fault detection method based on adaptive continuous sampling, such as Figure 1 As shown, the specific steps include:

[0034] Step (1): Select the sensor type, determine the sensor installation location and detection parameters;

[0035] Step (2): Acquire synchronously collected vibration data;

[0036] Step (3): reconstructing synchronously collected vibration data;

[0037] Step (4): Calculate fault characteristics;

[0038] Step (5): Determine whether a fault has occurred and adaptively adjust the amount of synchronously collected data per rotation of the shaft for the next fault detection.

[0039] Furthermore, the step (1) of selecting the sensor type includes: the selected sensor is a vibration sensor and a rotation speed sensor.

[0040] Furthermore, the step (1) of determining the sensor installation position includes: the vibration sensor is installed near the measured bearing to measure the bearing vibration data; the speed sensor is installed near the shaft to measure the shaft speed signal.

[0041] Furthermore, the step (1) of determining the detection parameters includes: determining the initial amount N0, the minimum amount N1 and the maximum amount N2 of the synchronously collected data amount N per rotation of the shaft, the sampling revolution number k, and two thresholds T1 and T2; when the detected bearing is a brand new bearing, N0 is set to N min ; When the health status of the detected bearing is unknown, set Among them, round() is a rounding function; the value range of the sampling speed k is 5 to 20; the two thresholds T1 and T2 are determined according to the mean μ0 and standard deviation σ0 of the fault characteristics when the bearing is fault-free, T1 = μ0 and T2 = μ0 + 3σ0.

[0042] Further, the step (2) includes: based on the vibration sensor and the speed sensor, according to the synchronously collected data amount N per rotation of the shaft and the sampling rotation number k, obtaining the synchronously collected vibration data S of the bearing with a data length of kN, and the synchronously collected vibration signal S can be expressed as [x1, x2, x3, ..., x kN ].

[0043] Furthermore, the step (3) includes: reconstructing the one-dimensional synchronous vibration data S into N groups of one-dimensional synchronous vibration data Y = [S1, S2, S3, ..., S i ,…,S N ], where i = 1, 2, 3, ..., N, the i-th group of one-dimensional synchronous vibration data S i It can be expressed as [x i ,x (i+N) ,x (i+2N) ,...,x (i+(k-1)N) ].

[0044] Furthermore, the step (4) comprises: respectively calculating the standard deviation Z of N groups of one-dimensional synchronous vibration data Y = [Z1, Z2, Z3, ..., Z i ,…,Z N ], the standard deviation of the i-th group of one-dimensional synchronous vibration data is

[0045]

[0046] Among them, μ i is the mean of the i-th group of one-dimensional synchronous vibration data; then the mean μ of the standard deviation Z is calculated and used as the fault feature,

[0047]

[0048] Furthermore, the step (5) includes: judging whether the bearing fault occurs according to the obtained fault feature μ and the two set thresholds T1 and T2, and adaptively adjusting the amount of synchronously collected data N per rotation of the shaft for the next fault detection; when the fault feature μ < T1, it indicates that the bearing is in a healthy state, then the amount of synchronously collected data N per rotation of the shaft for the next fault detection is adjusted to the minimum value N1, and the process is turned to step (2) to continue the bearing fault detection process; when the fault feature T1 ≤ μ ≤ T2, it is impossible to reliably determine whether the bearing fault occurs, and at this time the amount of synchronously collected data N per rotation of the shaft for the next fault detection is adjusted to Among them, the function ceil() is a rounding up function, and goes to step (2) to continue the bearing fault detection process; when the fault feature μ>T2, it indicates that the bearing fault has occurred, and the bearing fault detection process is stopped.

[0049] In this embodiment, the initial amount N0 of synchronous sampling data per rotation of the shaft is set to 50, the minimum amount N1 is set to 1, and the maximum amount N2 is set to 100, the sampling rotation number k is set to 10, and the two thresholds T1 is set to 0.114 and T2 is set to 0.174.

[0050] The feasibility of the method of the present invention is tested based on the gearbox bearing fault simulation signal provided by the invention patent with application number 202410007749.2. The gearbox bearing fault simulation signal includes the following four components:

[0051] x(t)=x1(t)+x2(t)+x3(t)+x4(t)

[0052]

[0053] x2(t)=0.55sin(2πf2t)

[0054] x3(t)=sin(2πf3t)

[0055] x4(t)=0.4rand(n,1),n=length(t)

[0056] Wherein, x1(t) represents the modulation signal component of the bearing outer ring fault signal and the structural resonance signal at the large gear; x2(t) and x3(t) are periodic harmonic signal components; x4(t) is the simulated noise signal; α is the attenuation factor (related to the damping characteristics of the system), α = 400; f1 is the structural resonance frequency, f1 = 800 Hz; f2 is the meshing frequency of the high-speed gear, f2 = 180 Hz; f3 is the meshing frequency of the low-speed gear, f3 = 60 Hz, f B is the fault characteristic frequency, f B=33Hz; mod(·) is the modulo function; rand(·) is the random function; length(·) is the function for finding the length of an array.

[0057] The simulation signals when the gearbox bearing is fault-free and faulty are as follows: Figure 2 and Figure 3 As shown. According to the initial amount of synchronous sampling data N0 = 51 per rotation of the shaft, the speed sensor and the vibration sensor are used to synchronously collect vibration data S. The collected synchronously collected vibration data S is reconstructed into 51 groups of one-dimensional synchronous vibration data Y = [S1, S2, S3, ..., S i ,…,S 51 ], where i = 1, 2, 3, ..., 51. Calculate the standard deviation of all one-dimensional synchronous vibration data and the mean μ of these standard deviations, and regard the mean μ as the bearing fault feature. When the fault feature μ < T1, it indicates that the bearing is in a very healthy state, then the amount of synchronous sampling data N per rotation of the shaft for the next fault detection is adjusted to 1, and go to step 2 to continue the bearing fault detection process. When the fault feature T1 ≤ μ ≤ T2, it is impossible to reliably determine whether the bearing is faulty. At this time, the amount of synchronous sampling data N per rotation of the shaft for the next fault detection is adjusted to The function ceil() is a rounding up function, and the process goes to step 2 to continue the bearing fault detection process. When the fault feature μ>T2, it indicates that the bearing is faulty, and the bearing fault detection process is stopped.

[0058] Figure 4 and Figure 5 The detection results when the bearing is fault-free and the number of synchronously sampled data N per rotation of the shaft are shown respectively. In the 9 bearing fault detections, all fault features are located near the threshold T1 and are all less than the threshold T2, indicating that the bearing is not faulty. Except for the first bearing fault detection, the number of synchronously sampled data N per rotation of the shaft required for the remaining fault detections is much less than the maximum value of 100. The method of the present invention significantly reduces the amount of data collected per rotation of the shaft when the bearing is fault-free, reducing the cost of data transmission, processing and storage.

[0059] Figure 6 and Figure 7 The detection results of bearing failure and the number of synchronous sampling data N per shaft revolution are shown respectively. In the 9 bearing fault detections, all fault features are located near the threshold T2, and some fault features are greater than the threshold T2, indicating that the bearing has failed. Except for the first bearing fault detection, the number of synchronous sampling data N per shaft revolution required for each fault detection is approximately a maximum of 100. When the bearing fails, the method of the present invention increases the amount of data collected per shaft revolution cycle, ensuring the reliability of bearing fault detection.

[0060] The above description is only an example of the present invention and is not intended to limit the present invention. Any equivalent substitution made within the principle of the present invention should be included in the protection scope of the present invention. The content not elaborated in detail in the present invention belongs to the existing technology known to the technicians in this professional field.

Claims

1. A bearing fault detection method based on adaptive continuous sampling is characterized by: The following steps are involved: Step (1): Select the sensor type, determine the sensor installation location and detection parameters; Step (2): Acquire synchronously collected vibration data; Step (3): reconstructing synchronously collected vibration data; Step (4): Calculate fault characteristics; Step (5): Determine whether a fault has occurred and adaptively adjust the amount of synchronously collected data per rotation of the shaft for the next fault detection.

2. A bearing fault detection method based on adaptive continuous sampling according to claim 1, characterized in that: The step (1) of selecting the sensor type includes: the selected sensor is a vibration sensor and a rotation speed sensor.

3. A bearing fault detection method based on adaptive continuous sampling according to claim 2, characterized in that: The step (1) of determining the sensor installation position includes: the installation position of the vibration sensor is near the measured bearing; the installation position of the speed sensor is near the rotating shaft.

4. A bearing fault detection method based on adaptive continuous sampling according to claim 3, characterized in that: The step (1) of determining the detection parameters includes: determining the initial amount N0, the minimum amount N1 and the maximum amount N2 of the synchronous data collection amount N per rotation of the shaft, the sampling revolution number k, and two thresholds T1 and T2; when the detected bearing is a brand new bearing, N0 is set to N min ; When the health status of the detected bearing is unknown, set Among them, round() is a rounding function; the value range of the sampling speed k is 5 to 20; the two thresholds T1 and T2 are determined according to the mean μ0 and standard deviation σ0 of the fault characteristics when the bearing is fault-free, T1 = μ0 and T2 = μ0 + 3σ0.

5. A bearing fault detection method based on adaptive continuous sampling according to claim 4, characterized in that: The step (2) comprises: based on the vibration sensor and the speed sensor, according to the synchronously collected data amount N per rotation of the rotating shaft and the sampling rotation number k, obtaining the synchronously collected vibration data S of the bearing with a data length of kN, wherein the synchronously collected vibration data S can be expressed as [x1, x2, x3, ..., x kN ].

6. A bearing fault detection method based on adaptive continuous sampling according to claim 5, characterized in that: The step (3) comprises: reconstructing the one-dimensional synchronous vibration data S into N groups of one-dimensional synchronous vibration data Y=[S1, S2, S3, ..., S i ,…,S N ], where i = 1, 2, 3, ..., N, the i-th group of one-dimensional synchronous vibration data S i It can be expressed as [x i ,x (i+N) ,x (i+2N ),...,x (i+(k-1)N) ].

7. A bearing fault detection method based on adaptive continuous sampling according to claim 6, characterized in that: The step (4) comprises: respectively calculating the standard deviation Z of N groups of one-dimensional synchronous vibration data Y = [Z1, Z2, Z3, ..., Z i ,…,Z N ], the standard deviation of the i-th group of one-dimensional synchronous vibration data is Among them, μ i is the mean of the i-th group of one-dimensional synchronous vibration data; then the mean μ of the standard deviation Z is calculated and used as the fault feature, 8. A bearing fault detection method based on adaptive continuous sampling according to claim 7, characterized in that: The step (5) comprises: judging whether the bearing fault occurs according to the obtained fault feature μ and the two set threshold values ​​T1 and T2, and adaptively adjusting the amount of synchronously collected data N per rotation of the shaft for the next fault detection; when the fault feature μ < T1, it indicates that the bearing is in a healthy state, then the amount of synchronously collected data N per rotation of the shaft for the next fault detection is adjusted to a minimum value N1, and the process is turned to step (2) to continue the bearing fault detection process; when the fault feature T1 ≤ μ ≤ T2, it is impossible to reliably determine whether the bearing fault occurs, and at this time the amount of synchronously collected data N per rotation of the shaft for the next fault detection is adjusted to Among them, the function ceil() is a rounding function, and goes to step (2) to continue the bearing fault detection process; when the fault feature μ>T2, it indicates that the bearing fault has occurred, and the bearing fault detection process is stopped.

Images