Intelligent bearing fault position resolving method fusing vibration and strain signals

Abstract

According to the intelligent bearing fault position calculation method, vibration and strain signals are fused, a rolling bearing retainer rotating speed measurement algorithm based on the strain signals and a calculation order tracking algorithm with the retainer as a reference axis are constructed, and accurate fault position calculation under the variable rotating speed condition, especially in the slipping state is achieved. According to the method, the rotating speed of the bearing retainer is measured innovatively by using the strain signal, and the problem of fault characteristic frequency deviation caused by slipping is effectively solved. According to the method, the fault characteristic vibration and order can still be accurately corresponding under the conditions of rotating speed change and slipping, and the precision and reliability of fault position calculation are improved. Compared with a traditional order tracking method, the envelope order spectrum generated by the method has no order confusion, the fault feature order is clearly indicated, and a more efficient and accurate solution is provided for bearing fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical fields of mechanical fault diagnosis technology and signal processing and analysis technology, and particularly relates to an intelligent bearing fault position calculation method that fuses vibration and strain signals. Background Art

[0002] Vibration analysis is widely used in bearing fault diagnosis due to its sensitivity to early bearing faults and ease of use. When the bearing is operating at variable speeds, the vibration signal has significant non-stationary characteristics, manifested as frequency ambiguity in the frequency spectrum, which makes it difficult to effectively apply traditional signal processing methods. The order tracking technology can effectively eliminate the influence of speed changes on the signal by resampling the vibration signal in the angular domain, accurately identify the fault characteristic components, and thus determine the bearing fault position.

[0003] The patent with the patent application number CN201811250171.4 uses the calculation order tracking algorithm to diagnose the faults of rolling bearings under variable speed conditions. This method synchronously collects vibration and spindle speed signals, obtains the instantaneous phase of the bearing inner ring through the spindle speed signal, and then uses this phase information to resample the fault impact envelope in the angular domain, realizing the bearing fault diagnosis under variable speed conditions. During the actual operation of the bearing, there is a phenomenon of cage slippage, which is particularly serious when the bearing is under high-speed light load and variable speed conditions. When there is cage slippage, using the spindle speed signal for angular resampling of fault impacts will still have the problem of frequency ambiguity, and the impact order in the signal cannot correspond to the theoretical fault order, making it impossible to accurately calculate the bearing fault position.

[0004] The intelligent bearing can sense the pulses generated by the rollers passing through the measuring point in sequence by arranging strain sensitive elements on the rolling bearing rings, thereby measuring the cage speed in-situ. Establishing an intelligent bearing fault position calculation method that fuses vibration and strain signals can realize the order tracking of the vibration signal of the rolling bearing and fault diagnosis under the slipping state. Summary of the Invention

[0005] The present invention proposes an intelligent bearing fault position calculation method, which fuses vibration and strain signals, constructs a rolling bearing cage speed measurement based on strain signals and a calculation order tracking algorithm with the cage as the reference axis, and realizes the calculation of the fault position of the rolling bearing under variable speed conditions, especially under the slipping state. The research results show that this method can accurately measure the cage speed of the rolling bearing and has a high recognition ability for the rolling bearing fault position.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] Step 1: Groove the intelligent bearing race, arrange the strain-sensitive element in the embedded groove, install the vibration sensor on the bearing housing, and install the tachometer on one side of the main shaft. Synchronously collect the strain signal δ(t), vibration signal x(t), and main shaft tachometer signal v(t) through the data acquisition system;

[0008] Step 2: Perform noise reduction processing on the strain signal by sliding average and detrending. Use the set threshold to extract the rising edge moment of the waveform, and interpolate to obtain the cage instantaneous phase signal θ δ (t), and differentiate to obtain the cage instantaneous rotational speed signal s δ (t). Process the main shaft tachometer signal v(t) to obtain the main shaft phase signal θ v (t) and the instantaneous rotational speed signal s v (t);

[0009] Step 3: Determine the frequency band where the fault impact component is located in the vibration signal x(t) by the Kurtogram method. Perform band-pass filtering and Hilbert transform on this frequency band to extract the fault impact envelope x e (t);

[0010] Step 4: Use the calculation order tracking algorithm to perform order tracking on the fault impact envelope x e (t). Perform two types of angular domain resampling according to the possible fault positions of the bearing. Specifically: for possible outer ring, cage, and roller position faults, use the cage as the reference axis, and through the cage instantaneous phase signal θ δ (t), perform angular domain resampling on the fault impact envelope x e (t), and the angular domain resampled signal is x e (θ δ ); for possible inner ring position faults, use the difference between the shaft and the cage rotation frequency as the reference, and through the difference between the main shaft and cage instantaneous phases θ v (t)-θ δ (t), perform angular domain resampling on the fault impact envelope x e (t), and the angular domain resampled signal is x e (θ v-δ );

[0011] Step 5: Calculate the bearing fault characteristic order. The calculation of the traditional bearing fault characteristic order is based on the pure rolling assumption. However, the present invention aims to explore the fault diagnosis of the rolling bearing cage in the case of slipping. Therefore, the concept of the fault characteristic order is re-interpreted and defined:

[0012] The cage slip rate η of the bearing is the ratio of the actual rotational speed of the cage to the rotational speed under the pure rolling assumption, and is obtained from equation (1):

[0013]

[0014] The outer race fault characteristic order BPOO (Ball Pass Order Outer Race) is referenced to the cage rotation frequency:

[0015] BPOO = n (2)

[0016] The inner race fault characteristic order BPOI (Ball Pass Order Inner Race) is referenced to the difference between the shaft rotation frequency and the cage rotation frequency:

[0017] BPOI = n (3)

[0018] The rolling element fault characteristic order BSO (Ball Spin Order), calculating BSO with the cage rotation frequency as the reference and considering the influence of slip, can be expressed by the following formula according to the linear velocity relationship at the contact position between the rolling element and the inner race:

[0019] f bsf d + f m (D m - dcosα) = f i (D m - dcosα)η (4)

[0020] Substituting Equation (1) into Equation (4), we can get:

[0021]

[0022] Thus, we obtain:

[0023]

[0024] The cage fault characteristic order FTO (Cage Fault Order) is referenced to the cage rotation, and FTO is 1;

[0025] In the formula, f m is the actual rotation frequency of the cage, f FTF is the cage rotation frequency calculated from the rotation speed of the rotating shaft under the pure rolling assumption, f i is the inner race rotation frequency, f bsf is the self-rotation frequency of the rolling element, n is the number of rolling elements, D m is the pitch diameter of the raceway, d is the diameter of the rolling element, and α is the contact angle;

[0026] Step 6: Perform Fourier transform on the fault impact envelopes x e (θ δ ) and x e (θ v-δ ) to obtain the envelope order spectrum, x e (θδ ) If the characteristic order harmonics of BPOO (outer raceway position fault), FTO (cage position fault), or BSO (rolling element position fault) appear in the envelope order spectrum, it indicates that there is a fault at the corresponding position, x e (θ v-δ ) If the characteristic order harmonics of BPOI appear in the envelope order spectrum, it indicates that there is a fault at the inner raceway position.

[0027] The present invention has the following beneficial effects:

[0028] a) The method of the present invention innovatively collects strain signals to measure the rotational speed of the bearing cage, can accurately calculate the bearing fault position under the condition of slip, and overcomes the problem of the characteristic frequency shift of the fault caused by slip;

[0029] b) The bearing fault characteristic order is a constant, independent of the rotational frequencies of the bearing and the cage, and can accurately correspond the characteristic vibration caused by the fault to the order even when the rotational speed changes and there is slip, which is beneficial to the integration and application of the fault position calculation algorithm;

[0030] c) When there is slip in the bearing, the spectrogram obtained by the traditional order tracking method often has a certain order ambiguity problem. The envelope order spectrum obtained by the method of the present invention has no order confusion and can clearly indicate the fault characteristic order. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a schematic diagram of the present invention.

[0032] Figure 2 is a photo of the test bench of the embodiment of the present invention.

[0033] Figure 3 is a flowchart of the present invention.

[0034] Figure 4 is the strain signal δ(t) collected by the strain gauge in the embodiment of the present invention.

[0035] Figure 5 is the vibration signal x(t) collected by the vibration sensor in the embodiment of the present invention.

[0036] Figure 6 is the spindle tachometer signal v(t) collected by the spindle tachometer in the embodiment of the present invention.

[0037] Figure 7 is a comparison diagram of the instantaneous phase of the cage obtained by the spindle tachometer and the instantaneous phase of the cage obtained by the strain signal in the embodiment of the present invention.

[0038] Figure 8 is a comparison diagram of the instantaneous rotational speed of the cage obtained by the spindle tachometer and the instantaneous rotational speed of the cage obtained by the strain signal in the embodiment of the present invention.

[0039] Figure 9 For the fault impact envelope x of the embodiment of the present invention e (t) spectrogram.

[0040] Figure 10 For the envelope order spectrum obtained by order tracking of the strain signal in the embodiment of the present invention.

[0041] Figure 11 For the envelope order spectrum obtained by order tracking of the spindle tachometer signal in the embodiment of the present invention. Detailed implementation manners

[0042] The present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0043] As Figure 1 is a schematic diagram of this embodiment. 1 is a bearing housing, 2 is a strain gauge and the grooving position of the bearing ring, 3 is a vibration sensor, 4 is a vibration signal, 5 is a strain signal, 6 is a rotational speed signal, 7 is a data acquisition card for synchronously acquiring the three signals, and 8 is a host computer for analysis algorithms.

[0044] The specific parameters are as follows: 1) The bearing inside the bearing housing 1 is an outer ring fault bearing with a fault width of 0.2 mm; 2) The spindle speed is driven by a servo motor to increase from 0 RPM to 1000 PRM and then decrease to 0 RPM within 20 seconds in a sinusoidal speed curve, no load; 3) The fault bearing model is SKF 6306 deep groove ball bearing, and the specific parameters are: a) Roller diameter: 12.3 mm; b) Pitch diameter of the raceway: 52 mm; c) Number of rollers: 8; d) Contact angle: 0°; 4) Sampling frequency is 50000 Hz; 5) Through Figure 2 The tachometer signal, strain signal and vibration signal obtained by conducting experiments on the bearing test bench verify the effectiveness of the present invention.

[0045] Apply the present invention to perform fusion analysis on the vibration signal and the strain signal and calculate the bearing fault position. As Figure 3 shown is the flow chart of the inventive method, including the following steps:

[0046] Step 1: Groove the intelligent bearing ring, arrange the strain sensitive element in the embedded groove, install the vibration sensor on the bearing housing, install the tachometer on one side of the spindle, and synchronously acquire the strain signal δ(t) ( Figure 4 ), vibration signal x(t) ( Figure 5 ) and spindle tachometer signal v(t) ( Figure 6 ) through the data acquisition system;

[0047] Step 2: Perform denoising processing of moving average and detrending on the strain signal, extract the rising edge moment of the waveform using a set threshold, and interpolate to obtain the cage instantaneous phase signal θδ (t), differentiating to obtain the cage instantaneous rotational speed signal s δ (t), processing the spindle tachometer signal v(t) to obtain the spindle phase signal θ v (t) and the instantaneous rotational speed signal s v (t), comparing the actual instantaneous phase signal of the cage obtained from the strain signal and the instantaneous phase signal of the cage obtained from the spindle tachometer as Figure 7 shown, the actual instantaneous phase of the cage lags behind the instantaneous phase of the cage obtained by converting the spindle phase, and the comparison of the actual instantaneous rotational speed of the cage obtained from the strain signal and the instantaneous rotational speed of the cage obtained by converting the spindle rotational speed is as Figure 8 shown, it can be shown that there is slippage of the cage;

[0048] Step Three: Determine the frequency band where the fault impact component is located in the vibration signal x(t) by the spectral kurtosis (Kurtogram) method, perform band-pass filtering and Hilbert transform on this frequency band to extract the fault impact envelope x e (t), the direct spectrum analysis result is as Figure 9 shown, there is a problem of frequency ambiguity caused by variable rotational speed;

[0049] Step Four: Adopt the computational order tracking algorithm to perform order tracking on the fault impact envelope x e (t), perform two types of angular domain resampling according to the possible fault positions of the bearing, specifically: for possible outer ring, cage and roller position faults, taking the cage as the reference axis, through the cage instantaneous phase signal θ δ (t) perform angular domain resampling on the fault impact envelope x e (t), the angular domain resampled signal is x e (θ δ ); for possible inner ring position faults, taking the difference between the shaft and the cage rotation frequency as the reference, through the difference between the spindle and cage instantaneous phases θ v (t) - θ δ (t) perform angular domain resampling on the fault impact envelope x e (t), the angular domain resampled signal is x e (θ v-δ ), in this embodiment, the fault sample is an outer ring fault and the reference axis is the cage;

[0050] Step Five: Calculate the bearing fault characteristic order. The calculation of the traditional bearing fault characteristic order is based on the pure rolling assumption, while the present invention aims to explore the fault diagnosis of the rolling bearing cage in the case of slippage. Therefore, the concept of the fault characteristic order is re-interpreted and defined:

[0051] The slip rate η of the bearing cage is the ratio of the actual rotational speed of the cage to the rotational speed under the pure rolling assumption, and is obtained from Equation (1):

[0052]

[0053] The outer race fault characteristic order BPOO (Ball Pass Order Outer Race) is referenced to the cage rotation frequency:

[0054] BPOO = n (2)

[0055] The inner race fault characteristic order BPOI (Ball Pass Order Inner Race) is referenced to the difference between the shaft rotation frequency and the cage rotation frequency:

[0056] BPOI = n (3)

[0057] The rolling element fault characteristic order BSO (Ball Spin Order), with the cage rotation frequency as a reference to calculate BSO, considering the influence of slip, can be expressed by the following formula based on the linear velocity relationship at the contact position between the rolling element and the inner race:

[0058] f bsf d + f m (D m - dcosα) = f i (D m (D - dcosα)η (4)

[0059] Substituting Equation (1) into Equation (4), we can obtain:

[0060]

[0061] From this, we get:

[0062]

[0063] The cage fault characteristic order FTO (Cage Fault Order) is referenced to the cage rotation, and FTO is 1;

[0064] In the formula, f m is the actual rotational frequency of the cage, f FTF is the rotational frequency of the cage calculated from the rotational speed of the rotating shaft under the pure rolling assumption, f i is the inner race rotational frequency, f bsf is the rotational frequency of the rolling element self - rotation, n is the number of rolling elements, D m is the pitch diameter of the raceway, d is the diameter of the rolling element, and α is the contact angle;

[0065] In this embodiment, it is an outer race fault, the reference axis is the cage, and the fault order is 8;

[0066] Step 6: The fault impact envelope x e (θ δ ) and x e (θ v-δ ) are Fourier-transformed to obtain the envelope order spectrum. If characteristic order harmonics of BPOO (outer raceway position fault), FTO (cage position fault), or BSO (rolling element position fault) appear in the envelope order spectrum of x e (θ δ ), it indicates that there is a fault at the corresponding position. If characteristic order harmonics of BPOI appear in the envelope order spectrum of x e (θ v-δ ), it indicates that there is an inner raceway position fault;

[0067] The envelope order spectrum obtained in this embodiment is as shown in Figure 10 . The spectrogram clearly indicates the characteristic order 8 of the outer raceway fault, effectively diagnosing the existing outer raceway fault. In contrast, the result of the envelope order spectrum obtained by order tracking using the spindle tachometer signal is as shown in Figure 11 . Although the result reduces the frequency ambiguity caused by speed changes to a certain extent, there is still a problem of ambiguity caused by slipping, and the characteristic order shifts from 8 to 7.53, unable to clearly indicate the outer raceway fault;

[0068] The intelligent bearing fault position calculation method based on the fusion of vibration and strain signals proposed by the present invention can accurately obtain the instantaneous phase signal of the cage even under the bearing slipping state, thereby performing equal-angle domain resampling on the fault impact envelope, extracting the fault position information, and accurately calculating the fault position of the rolling bearing with slipping under variable speed conditions.

Claims

1. An intelligent bearing fault location calculation method that fuses vibration and strain signals, characterized in that, The method includes the following steps: Step 1: Groove the intelligent bearing ring, arrange the strain sensitive element in the embedded groove, install the vibration sensor on the bearing housing, and install the tachometer on one side of the main shaft. Synchronously collect the strain signal δ(t), vibration signal x(t) and main shaft tachometer signal v(t) through the data acquisition system; Step 2: Perform noise reduction processing of sliding average and detrending on the strain signal, extract the rising edge moment of the waveform using a set threshold, and interpolate to obtain the instantaneous phase signal θ δ (t), and differentiate to obtain the instantaneous rotational speed signal s δ (t). Process the spindle tachometer signal v(t) to obtain the spindle phase signal θ v (t) and the instantaneous rotational speed signal s v (t); Step 3: Determine the frequency band where the fault impact component in the vibration signal x(t) is located by the Kurtogram method, and perform band-pass filtering and Hilbert transform on this frequency band to extract the fault impact envelope x e (t); Step 4: Perform order tracking on the fault impact envelope x e (t) using the computational order tracking algorithm; Step 5: Calculate the bearing fault characteristic order; Step 6: The fault impact envelope x e (θ δ ) and x e (θ v-δ ) are subjected to Fourier transform to obtain the envelope order spectrum. If the characteristic order harmonics of BPOO (outer raceway position fault), FTO (cage position fault), or BSO (rolling element position fault) appear in the envelope order spectrum of x e (θ δ ), it indicates that there is a fault at the corresponding position. If the characteristic order harmonics of BPOI appear in the envelope order spectrum of x e (θ v-δ ), it indicates that there is a fault at the inner raceway position.

2. The intelligent bearing fault location calculation method for fusing vibration and strain signals according to claim 1, characterized in that: The order tracking algorithm is used to perform order tracking on the fault impact envelope x e (t). Two types of angular domain resampling are carried out according to the possible fault positions of the bearing. Specifically: for possible outer ring, cage, and roller position faults, with the cage as the reference axis, the fault impact envelope x δ (t) is resampled in the angular domain through the instantaneous phase signal θ e (t) of the cage, and the angular domain resampled signal is x e (θ δ ); for possible inner ring position faults, with the difference between the shaft and cage rotation frequencies as the reference, the fault impact envelope x v (t) is resampled in the angular domain through the difference between the instantaneous phases of the main shaft and the cage θ δ (t)-θ e (t), and the angular domain resampled signal is x e (θ v-δ ).

3. The intelligent bearing fault location calculation method for fusing vibration and strain signals according to claim 1, wherein: Calculate the bearing fault characteristic order, and re-interpret and define the concept of the bearing fault characteristic order for the case of the rolling bearing cage slipping: The cage slip rate η of the bearing is the ratio of the actual speed of the cage to the speed under the pure rolling assumption, and is obtained from equation (1): The outer race fault characteristic order BPOO (Ball Pass Order Outer Race) is referenced to the cage rotation frequency: BPOO = n (2) The inner race fault characteristic order BPOI (Ball Pass Order Inner Race) is referenced to the difference between the shaft and cage rotation frequencies: BPOI = n (3) The rolling element fault characteristic order BSO (Ball Spin Order), calculate BSO with reference to the cage rotation frequency, considering the influence of slip, and can be expressed by the following formula according to the linear velocity relationship at the contact position between the rolling element and the inner race: f bsf d + f m (D m - dcosα) = f i (D m - dcosα)η (4) Substitute equation (1) into equation (4), and we can get: Thus, we obtain: The cage fault order FTO (Cage Fault Order) is referenced to the rotation of the cage, and FTO is 1; where f m is the actual rotational frequency of the cage, f FTF is the rotational frequency of the cage calculated from the rotational speed of the rotating shaft under the pure rolling assumption, f i is the rotational frequency of the inner ring, f bsf is the rotational frequency of the rolling element, n is the number of rolling elements, D m is the pitch diameter of the raceway, d is the diameter of the rolling element, and α is the contact angle.

Images