Rolling bearing early fault real-time detection method based on EWMA

Abstract

The invention discloses an EWMA-based rolling bearing early fault real-time detection method. The method comprises the steps of setting mounting positions and detection parameters of a vibration sensor and a rotating speed sensor; acquiring each group of period synchronous single-point sampling data in the current time window; calculating standard deviations of the synchronous single-point sampling data of each group of periods in the current time window, and calculating an average value of the standard deviations to obtain fault features; accumulating fault information contained in the fault features by using an exponential weighted moving average (EWMA) method to obtain a bearing health index; judging whether an early fault occurs or not according to the bearing health index, if no fault occurs, adaptively adjusting the synchronous sampling data volume in each rotation period of the rotating shaft, and returning to the step 2 to continue the early fault detection of the bearing in the next time window; according to the invention, the synchronous sampling data volume in each rotation period of the rotating shaft is adaptively determined according to the bearing health index, the data acquisition amount is reduced under the condition that the fault detection reliability and sensitivity are ensured, and the data transmission, processing and storage costs are further reduced.

Description

Technical Field

[0001] The invention relates to a real-time detection method for early-stage faults of rolling bearings, and in particular to a real-time detection method for early-stage faults of rolling bearings based on EWMA, and belongs to the technical field of rotating machinery fault detection and diagnosis. Background Art

[0002] Rotating machinery is a type of mechanical equipment that achieves its working purpose through rotational motion. They have been widely used in many fields such as industry, transportation, and energy. Common rotating machinery includes: motors, generators, turbines, compressors, etc. Rotating machinery may have various faults during long-term operation. Bearing failure is the most common form of failure in rotating machinery. In order to ensure the stable operation of rotating machinery and extend its service life, real-time fault detection and regular maintenance of bearings are required.

[0003] At present, the real-time detection technology for early bearing faults usually adopts a fixed sampling frequency strategy for data collection. The fixed sampling frequency strategy does not take into account the degradation of the equipment and environmental changes, and therefore lacks flexibility in data sampling. The collected data obtained by the fixed sampling frequency may have problems such as data redundancy or insufficiency. When data redundancy occurs, the cost of data transmission, processing and storage will increase, thereby reducing the sensitivity of fault detection, and when the collected data is insufficient, the reliability of fault detection will decrease. In addition, due to the influence of disturbance noise, weak fault information is difficult to extract, so the sensitivity and reliability of early bearing fault detection are relatively low. In order to solve the above problems, the patent of this invention proposes a real-time detection method for early rolling bearing faults based on EWMA. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a real-time detection method for early rolling bearing faults based on EWMA, which can reduce the amount of sampling data while ensuring the sensitivity and reliability of fault detection.

[0005] Technical solution: The present invention discloses a real-time detection method for early failure of rolling bearings based on EWMA, comprising:

[0006] Step 1: Set the installation position and detection parameters of the vibration sensor and speed sensor;

[0007] Step 2: Obtain the synchronous single-point sampling data of each group of periods within the current time window;

[0008] Step 3: Calculate the standard deviation of each group of periodic synchronous single-point sampling data in the current time window, and calculate the average value of the standard deviation to obtain the fault characteristics;

[0009] Step 4: Using the exponentially weighted moving average (EWMA) method, the fault information contained in the fault characteristics is accumulated to obtain the bearing health index;

[0010] Step 5: Determine whether an early fault occurs based on the bearing health index. If there is no fault, adaptively adjust the amount of synchronous sampling data within each rotation period of the shaft, and return to step 2 to continue early bearing fault detection in the next time window.

[0011] Furthermore, setting the installation positions of the vibration sensor and the rotation speed sensor in step 1 includes installing the vibration sensor at the bearing to be tested to collect the bearing vibration signal; and installing the rotation speed sensor near the shaft to measure the shaft rotation speed.

[0012] Furthermore, the setting of detection parameters in step 1 includes:

[0013] Set the initial value and minimum value N of the number of synchronously sampled data points per rotation period of the shaft min and the maximum value N max ; Set the shaft speed L for data collection.

[0014] Furthermore, the initial value of the number N of synchronously sampled data points per rotation period of the shaft is specifically set as follows: when the bearing to be tested is a brand new bearing, N is set to N min , when the health status of the detected bearing is unknown, set N= Among them, round() is a rounding function, and the result is an integer.

[0015] Furthermore, setting the detection parameters in step 1 also includes setting the upper control limit of EWMA to threshold T1, setting the lower control limit to threshold T2, setting the forgetting factor λ to a value within (0, 1), and setting the average value of the standard deviation of the periodically synchronized single-point sampling data when the bearing is fault-free to mean M0.

[0016] Further, the step 2 comprises:

[0017] According to the number of synchronous sampling data points N in each rotation period of the shaft, obtain the N groups of periodic synchronous single-point sampling data Y in the current kth time window k , N groups of periodic synchronous single-point sampling data Y in the kth time window k It is expressed as:

[0018]

[0019] Among them, Y k The nth row of data in the matrix represents the nth group of periodic synchronous single-point sampling data in the kth time window, where n=1, 2, 3, …, N.

[0020] Furthermore, step 3 includes respectively calculating the standard deviation S of the N groups of periodic synchronous single-point sampling data in the current k-th time window k =[S k (1)S k (2)S k (3)…S k (n)…S k (N)] T , where the standard deviation of the nth group of periodic synchronous single-point sampling data in the kth time window is:

[0021]

[0022] Where: μ k (n) is the mean value of the nth group of periodic synchronous single-point sampling data,

[0023]

[0024] Then we get N standard deviations S in the kth time window k The average value Z k , and consider it as a fault feature:

[0025]

[0026] Furthermore, step 4 includes, according to the EWMA method, accumulating the fault characteristics Z k Contains the fault information and obtains the bearing health index X of the current k-th time window k :

[0027] X k =λ·Z k +(1-λ)·X k-1

[0028] Among them, X k-1 is the bearing health index of the k-1th time window. When detecting in the first time window, X0 is equal to the preset mean M0.

[0029] Further, the step 5 includes: k Compare with threshold T1; when X k If it is less than T1, it is determined that there is no fault, and X k =max(X k , T2), the number of synchronous sampling data points N in each rotation period of the shaft during the next time window detection is adaptively adjusted to And return to step 2 to continue the bearing early fault detection in the next time window; when X k If it is greater than or equal to T1, it is considered that an early failure has occurred.

[0030] Furthermore, the shaft speed L has a value range of 5 to 10.

[0031] Beneficial effects: Compared with the prior art, the present invention overcomes the shortcomings of existing real-time bearing fault detection methods, can adaptively determine the number of synchronously sampled data points within each rotation period of the shaft, and effectively reduce the amount of data collection, thereby reducing data transmission, processing and storage costs, and improving data transmission and processing efficiency; the method of the present invention utilizes EWMA technology to accumulate multiple time windows to adaptively obtain weak fault information in the data, thereby improving the reliability and sensitivity of early fault detection; the computational complexity of the method of the present invention is low, and a control unit with a lower processing speed can be used, thereby reducing the cost of the control unit; the method of the present invention is not affected by the rotation speed of the rotating machinery. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram of the detection result when the bearing is fault-free according to an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the number of synchronously sampled data points in each rotation period of the shaft when there is no bearing failure in an embodiment of the present invention;

[0035] Figure 4 This is a schematic diagram of the detection result of an early bearing failure according to an embodiment of the present invention;

[0036] Figure 5 This is a schematic diagram of the number of synchronously sampled data points in each rotation period of the shaft when a bearing fails in an early stage according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The implementation process of the method of the present invention is described below with reference to the accompanying drawings.

[0038] As attached Figure 1 As shown, the real-time detection method for early rolling bearing faults based on EWMA in this embodiment includes:

[0039] Step 1: Set the installation position and detection parameters of the vibration sensor and speed sensor;

[0040] Step 2: Obtain the synchronous single-point sampling data of each group of periods within the current time window;

[0041] Step 3: Calculate the standard deviation of each group of periodic synchronous single-point sampling data in the current time window, and calculate the average value of the standard deviation to obtain the fault characteristics;

[0042] Step 4: Using the exponentially weighted moving average (EWMA) method, the fault information contained in the fault characteristics is accumulated to obtain the bearing health index;

[0043] Step 5: Determine whether an early fault occurs based on the bearing health index. If there is no fault, adaptively adjust the amount of synchronous sampling data within each rotation period of the shaft, and return to step 2 to continue early bearing fault detection in the next time window.

[0044] Specifically, in step 1, a vibration sensor is installed at the pre-detection bearing to collect bearing vibration signals; a speed sensor is installed near the shaft to measure the shaft speed; the initial number of synchronously sampled data points per rotation period of the shaft is determined according to the health status of the detected bearing; in this embodiment, the health status of the measured bearing is unknown, and the initial number N0 of synchronously sampled data points per rotation period of the shaft is set to Among them, round() is a rounding function, and the result is an integer; the minimum number of synchronous sampling data points N in each rotation period of the shaft min =10 and the maximum number N max =100; therefore, the initial number of synchronously sampled data points N0=55 in each rotation period of the shaft; set the shaft rotation number L=10, the forgetting factor λ=0.9, the mean M0=0.3319, the upper threshold T1=0.3634, and the lower threshold T2=0.3004.

[0045] Step 2: According to the number of synchronous sampling data points N in each rotation period of the shaft set adaptively in the k-1th time window, obtain N groups of periodic synchronous single-point sampling data Y in the kth time window k Since the health status of the measured bearing is unknown, the initial number of synchronous sampling data points N0 per rotation period of the shaft is set to For the first time window, 55 sets of periodic synchronous single-point sampling data Y1 can be obtained, which can be expressed as:

[0046]

[0047] Each row of data in the Y1 matrix represents a set of periodic synchronous single-point sampling data in the first time window. Since the number of rotations of the shaft for data acquisition is L=10, the length of each set of periodic synchronous single-point sampling data is 10.

[0048] Specifically, step 3: calculate the standard deviation S of each group of periodic synchronous single-point sampling data in the kth time window respectively k , and then we can get the average value Z of N standard deviations in the kth time window k , which is used as the fault feature. For the first time window, the standard deviation of 55 sets of periodic synchronous single-point sampling data is

[0049] S1=[S1(1) S1(2) S1(3) … S1(n) … S1(55)] T ,

[0050]

[0051] Where: μ1(n) is the mean value of the nth group of periodic synchronous single-point sampling data in the first time window,

[0052] Then we get the mean Z1 of 55 standard deviations S1 in the first time window:

[0053]

[0054] This is considered a fault feature.

[0055] Specifically, step 4: According to the EWMA method, accumulate the fault characteristics Z in step 3 k Contains fault information and obtains the bearing health index X k .

[0056] For the first time window, the EWMA method is used to obtain the bearing health index X1

[0057] X1=0.9·Z1+(1-0.9)·X0,

[0058] Among them, X0=M0=0.3319, where M0 is the preset mean.

[0059] Specifically, step 5: convert the bearing health index X k Compare with threshold T1 = 0.3634. If the fault indicator X k If it is less than T1, it is determined that there is no fault, and X k =max(X k , T2), the number of synchronous sampling data points N in each rotation period of the shaft in the next time window is adaptively adjusted to And return to step 2 to continue the bearing early fault detection in the next time window; when X k If it is greater than or equal to the threshold T1, it is considered that an early fault has occurred.

[0060] Figure 2 is the detection result when the bearing is fault-free. As can be seen from the figure, the fault indicators are all less than the threshold T1, which means that the motor bearing is in a fault-free state. Let X k =max(X k , T2), the number of synchronous sampling data points N in each rotation period of the shaft detected in the next time window is adaptively adjusted to And return to step 2 to continue the bearing early fault detection in the next time window.

[0061] Figure 3The number of synchronously sampled data points N per rotation period of the shaft when the bearing is fault-free is shown. In each fault detection, the number of synchronously sampled data points per rotation period of the shaft changes with the fault index. Since the bearing is fault-free, the number of synchronously sampled data points N per rotation period of the shaft is less than the maximum value of 100 during each detection process, which reduces the amount of data collection and thus reduces the cost of data transmission, processing, and storage.

[0062] Figure 4 It is the detection result of the bearing failure. As can be seen from the figure, in the 9th time window, the fault index is greater than the threshold T1, which means that the bearing has an early fault.

[0063] Figure 5 is the number of synchronously sampled data points per rotation period of the shaft when the bearing has an early fault. In the 9th time window fault detection, the bearing has been detected as an early fault, and the early fault detection work is terminated, and the adaptive data acquisition is stopped.

[0064] The present invention patent discloses a real-time detection method for early rolling bearing faults based on EWMA. The present invention patent takes into account the degradation of the bearing and adaptively determines the number of sampling data points per rotation period of the shaft, that is, when the bearing degradation is relatively mild or no degradation, a smaller number of synchronous data is collected per rotation period of the shaft, and when the bearing degradation is more serious, a larger number of synchronous data is collected per rotation period of the shaft. Taking into account the degradation of the bearing, the present invention patent designs a fault detection method that adaptively determines the number N of synchronous sampling data points per rotation period of the shaft.

[0065] The method of the present invention accumulates weak fault information in the historical time window through the EWMA technology, which can make the early fault detection of the bearing more sensitive and reliable. In addition, compared with the traditional early fault detection method, the method of the present invention has a smaller computational complexity.

[0066] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A real-time detection method for early failure of rolling bearings based on EWMA, characterized in that: include: Step 1: Set the installation position and detection parameters of the vibration sensor and speed sensor; Step 2: Obtain the synchronous single-point sampling data of each group of periods within the current time window; Step 3: Calculate the standard deviation of each group of periodic synchronous single-point sampling data in the current time window, and calculate the average value of the standard deviation to obtain the fault characteristics; Step 4: Using the exponentially weighted moving average (EWMA) method, the fault information contained in the fault characteristics is accumulated to obtain the bearing health index; Step 5: Determine whether an early fault occurs based on the bearing health index. If there is no fault, adaptively adjust the amount of synchronous sampling data within each rotation period of the shaft, and return to step 2 to continue early bearing fault detection in the next time window.

2. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 1 is characterized in that: Setting the installation positions of the vibration sensor and the speed sensor in step 1 includes installing the vibration sensor at the bearing to be tested to collect the bearing vibration signal; and installing the speed sensor near the shaft to measure the shaft speed.

3. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 2 is characterized in that: Setting the detection parameters in step 1 includes: Set the initial value and minimum value N of the number of synchronously sampled data points per rotation period of the shaft min and the maximum value N max ; Set the shaft speed L for data collection.

4. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 3 is characterized in that: The initial value of the number N of synchronously sampled data points per rotation period of the shaft is specifically set as follows: when the bearing to be tested is a brand new bearing, N is set to N min , when the health status of the detected bearing is unknown, set Among them, round() is a rounding function, and the result is an integer.

5. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 4 is characterized in that: Setting the detection parameters described in step 1 also includes setting the upper control limit of EWMA to threshold T1, setting the lower control limit to threshold T2, setting the forgetting factor λ to a value within (0, 1), and setting the average value of the standard deviation of the periodic synchronous single-point sampling data when the bearing is fault-free to mean M0.

6. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 5 is characterized in that: The step 2 comprises: According to the number of synchronous sampling data points N in each rotation period of the shaft, obtain the N groups of periodic synchronous single-point sampling data Y in the current kth time window k , N groups of periodic synchronous single-point sampling data Y in the kth time window k It is expressed as: Among them, Y k The nth row of data in the matrix represents the nth group of periodic synchronous single-point sampling data in the kth time window, where n=1, 2, 3, …, N.

7. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 6 is characterized in that: The step 3 comprises respectively calculating the standard deviation S of the N groups of periodic synchronous single-point sampling data in the current k-th time window k =[S k (1)S k (2)S k (3)…S k (n)…S k (N)] T , where the standard deviation of the nth group of periodic synchronous single-point sampling data in the kth time window is: Where: μ k (n) is the mean value of the nth group of periodic synchronous single-point sampling data, Then we get N standard deviations S in the kth time window k The average value Z k , and consider it as a fault feature:

8. The real-time detection method for early rolling bearing faults based on EWMA according to claim 7 is characterized in that: The step 4 includes, according to the EWMA method, accumulating the fault characteristics Z k Contains the fault information and obtains the bearing health index X of the current k-th time window k : X k =λ·Z k +(1-λ)·X k-1 Among them, X k-1 is the bearing health index of the k-1th time window. When detecting in the first time window, X0 is equal to the preset mean M0.

9. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 8 is characterized in that: The step 5 includes: k Compare with threshold T1; when X k If it is less than T1, it is determined that there is no fault, and X k =max(X k , T2), the number of synchronous sampling data points N in each rotation period of the shaft during the next time window detection is adaptively adjusted to And return to step 2 to continue the bearing early fault detection in the next time window; when X k If it is greater than or equal to T1, it is considered that an early failure has occurred.

10. The real-time detection method for early failure of rolling bearing based on EWMA according to claim 3, characterized in that: The value range of the shaft speed L is 5 to 10.

Images