A noise power controllable acoustic fault diagnosis method for rolling bearings
By using a differential power beamformer based on second-order cone programming and an adaptive noise complete set empirical mode decomposition method, the problem of uncontrollable noise in the acoustic fault diagnosis of rolling bearings is solved, and fault identification and diagnosis with high signal-to-noise ratio are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-18
- Publication Date
- 2026-03-20
AI Technical Summary
Existing methods for diagnosing rolling bearing faults based on acoustic signals are easily affected by environmental noise, leading to uncontrollable noise, difficulties in signal processing, and challenges in accurately identifying the location and type of fault.
A differential power beamformer based on second-order cone programming is used, combined with a microphone array sensor and an adaptive noise complete set empirical mode decomposition method. The acoustic signal is processed by high- and low-power response beamformers to reduce noise and extract fault features.
It enables controllable noise power in the acoustic fault diagnosis of rolling bearings, improves the signal-to-noise ratio, accurately identifies the location and type of bearing faults, and ensures the safe and efficient operation of mechanical equipment.
Smart Images
Figure CN115901262B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of fault diagnosis and signal processing analysis, and particularly relates to a rolling bearing acoustic fault diagnosis method with controllable noise power. BACKGROUND
[0002] Rotating machinery (such as bearings) is the most common type of device in mechanical equipment and plays an important role in industrial applications, and is usually operated in harsh environments and prone to failure. Although the fault diagnosis method based on vibration signals is a mature and popular rolling bearing fault detection method, it has certain limitations in actual application due to the need for contact measurement. The acoustic-based diagnosis (ABD) technology has attracted widespread attention due to its non-contact nature, and thus, acoustic fault diagnosis can be regarded as an improvement over vibration signal-based diagnosis.
[0003] A microphone array can simultaneously obtain time-domain, spatial-domain and frequency-domain information of a target sound field. As a major task of array signal processing, the role of beamforming is to perform spatial filtering to collect sound signals from the target direction (bearing) without attenuation and to weaken sound signals from other directions (interference noise), which is exactly in line with the demand for obtaining high-quality acoustic signals. In the design of a beamformer, the size of the side lobe level is a key performance indicator that affects the quality of the output target direction signal. A low side lobe can more effectively suppress interference and noise from the side lobe region. However, if a sound signal with relatively high power appears in the side lobe region, the beamformer will lose its noise suppression effect. At this time, the noise is uncontrollable, causing signal processing difficulties and other problems. Therefore, it is of great significance to provide a new acoustic signal acquisition method that can obtain target bearing sound signals and power-controllable noise signals and accurately identify bearing fault positions and types for safe and efficient operation of mechanical equipment. SUMMARY
[0004] In view of the problem that the fault diagnosis method based on acoustic signals is prone to failure due to the influence of environmental noise, the application provides a rolling bearing acoustic fault diagnosis method with controllable noise power. Based on the difference power beam response method of the second-order cone programming, high signal-to-noise ratio rolling bearing acoustic signals are obtained, the noise power is controllable, and then the acoustic signals are processed to extract fault features, thereby realizing rolling bearing acoustic fault diagnosis.
[0005] To solve the above technical problems, the technical scheme adopted by the application is as follows: a rolling bearing acoustic fault diagnosis method with controllable noise power, comprising the following steps:
[0006] S1, collecting sound signals of the rolling bearing by using a microphone array sensor;
[0007] S2, establish a high-power response beamformer and a low-power response beamformer based on a second-order cone programming method;
[0008] S3, process the sound array signal through the high-power response beamformer and the low-power response beamformer, difference process the outputs of the two beamformers, and obtain a noise-controllable desired target acoustic signal;
[0009] S4, based on the desired target acoustic signal obtained in step S3, extract a fault feature and determine a fault type.
[0010] In the step S2, the specific method for establishing the high-power response beamformer and the low-power response beamformer based on the second-order cone programming method is as follows:
[0011] (1) initialize a sidelobe beam level, set a high desired beam response pattern and a low desired beam response pattern for a target azimuth angle; the response degree of the high desired beam response pattern at the target azimuth angle is higher than that of the low desired beam response pattern at the target azimuth angle;
[0012] (2) calculate the high-power response beamformer and the low-power beamformer based on a minimum sidelobe beam design method; the calculation formula is as follows:
[0013]
[0014] wherein w represents a weighting value vector of the beamformer, H represents a conjugate transpose, θ represents an azimuth angle, θ0 represents a main lobe direction, BW SL represents a beam sidelobe level, p(θ) represents a microphone array manifold vector, B d (θ) represents a beam response pattern.
[0015] In the step S4, the step of filtering the desired target acoustic signal is further included.
[0016] In the step S4, the adaptive noise complete set empirical mode decomposition method is used to filter the desired target acoustic signal.
[0017] In the step S4, after filtering the desired target acoustic signal, the step of feature enhancement of the desired target acoustic signal is further included.
[0018] In the step S4, the multi-point optimal minimum entropy deconvolution is used for signal feature enhancement.
[0019] In the step S4, the desired target acoustic signal is analyzed through envelope demodulation, a fault feature is extracted, and a fault type is determined.
[0020] Step S1 further includes analyzing the directional information of the target bearing sound source and noise source, and determining the optimal values for the microphone linear array spacing and the sound source distance.
[0021] Compared with the prior art, the present invention has the following advantages:
[0022] 1. This invention proposes a method for acoustic fault diagnosis of rolling bearings with controllable noise power. Based on a beamformer with target azimuth angle difference power response, the noise is spatially filtered and differentially processed using the main lobe as the power response reference, thereby obtaining a noise-controllable signal and improving the signal-to-noise ratio of the acoustic signal of the rolling bearing.
[0023] 2. The desired beamformer used in this invention is designed based on second-order cone optimization. Given the main lobe width, the beam is designed to effectively control the noise power of the beamformer output signal in the side lobe region.
[0024] 3. This invention combines spatial and temporal filtering methods to construct a differential power response beamformer to acquire signals, uses CEEMDAN-MOMEDA to reduce signal noise, and finally performs envelope demodulation analysis to complete bearing fault diagnosis, which can achieve more accurate fault diagnosis. Attached Figure Description
[0025] Figure 1 A flowchart illustrating a method for diagnosing acoustic faults in rolling bearings with controllable noise power, provided as an embodiment of the present invention.
[0026] Figure 2 This is a schematic diagram of the measurement principle in an embodiment of the present invention;
[0027] Figure 3 The target signal and its envelope spectrum constructed in the embodiments of the present invention;
[0028] Figure 4 The interference signal and its envelope spectrum constructed in the embodiments of the present invention;
[0029] Figure 5 This is a schematic diagram of the array signals captured in an embodiment of the present invention;
[0030] Figure 6 These are the high and low desired response beamformers calculated in the embodiments of the present invention;
[0031] Figure 7 This is a schematic diagram of the original signal and the signal after difference processing in an embodiment of the present invention;
[0032] Figure 8 To Figure 6 The adaptive noise-complete ensemble empirical mode decomposition results are obtained from the beam output difference signal in the middle.
[0033] Figure 9 The schematic diagram for the result of fault feature extraction. DETAILED DESCRIPTION
[0034] For the purpose, technical solutions and advantages of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.
[0035] As shown in the figure, the embodiment one of the present application provides a noise power controllable acoustic fault diagnosis method of rolling bearing, comprising: Figure 1
[0036] S1, collecting the sound signal of the rolling bearing by using the microphone array sensor.
[0037] In the step S1, the step of analyzing the azimuth information of the target bearing sound source and the noise source and determining the optimal value of the microphone linear array spacing and the sound source distance is further included. After setting the microphone array, the sound signal of the rolling bearing is collected by using the microphone array sensor, which can improve the amount of useful information carried by the signal collection. As shown in the figure, it is a measurement principle diagram of the present application, wherein S1: target sound source; Sn: interference noise source; φ: incident angle of the target sound source; d: microphone spacing; 1, 2, 3, …, N-1, N: microphone mark number; L: distance between the target sound source and the center of the microphone array. Figure 2
[0038] S2, establishing a high-power response beamformer and a low-power response beamformer based on the second-order cone programming method.
[0039] In the step S2, the specific method for establishing a high-power response beamformer and a low-power response beamformer based on the second-order cone programming method is:
[0040] (1) initializing the sidelobe beam level, setting a high expected beam response pattern and a low expected beam response pattern for the target azimuth angle; the response degree of the high expected beam response pattern at the target azimuth angle is higher than that of the low expected beam response pattern at the target azimuth angle.
[0041] (2) calculating the high-power response beamformer and the low-power beamformer based on the minimum sidelobe beam design method; the calculation formula is:
[0042]
[0043] Where w represents the weighted vector of the beamformer, H represents the conjugate transpose, θ represents the azimuth angle, θ0 represents the main lobe direction, and BW SL Let p(θ) represent the beamwidth of the beamside lobe, p(θ) represent the microphone array manifold vector, and B... d (θ) represents the beam response diagram, θ∈Θ SL Θ SL The sidelobe domain is represented. Substituting the high-expectation beam response map and the low-expectation beam response map into equation (1) respectively, the weighted vector w of the high-power response beamformer and the low-power beamformer can be obtained, that is, the corresponding high-power response beamformer and low-power beamformer can be obtained.
[0044] Where p is the microphone array manifold vector, which can be obtained by arranging the microphone array.
[0045] S3. The sound array signal is processed by a high-power response beamformer and a low-power response beamformer. The outputs of the two beamformers are differentially processed to obtain the desired target acoustic signal with controllable noise.
[0046] S4. Based on the desired target acoustic signal obtained in step S3, extract fault features and determine the fault type.
[0047] Step S4 also includes a step of filtering the desired target acoustic signal.
[0048] In step S4, the desired target acoustic signal is coarsely filtered using the adaptive noise complete set empirical mode decomposition method (CEEMDAN).
[0049] In step S4, after filtering the desired target acoustic signal, the method further includes a step of feature enhancement of the desired target acoustic signal. Specifically, multi-point optimal minimum entropy deconvolution is used to enhance the signal features.
[0050] Furthermore, in step S4, the desired target acoustic signal is analyzed by envelope demodulation to extract fault features and determine the fault type.
[0051] To demonstrate the effectiveness of this invention, according to Figure 2 The bearing fault acoustic signal model shown has the following key parameters: sampling frequency of 10kHz, signal length of 10000, characteristic frequencies of the target signal and interference signal of 130Hz and 190Hz respectively, angles between the signal incident angle and the microphone array axis of 30° and 45° respectively, and signal-to-noise ratio of the noise added to each signal of -10dB.
[0052] A simulation signal S1 containing bearing fault impact and Gaussian white noise is constructed as a target signal, and a simulation signal Sn containing bearing fault impact and Gaussian white noise is constructed as an interference signal, as shown in Figure 3 and Figure 4 Figure 3 and Figure 4 The original signal of the fault impact (a), the envelope spectrum of the original signal of the fault impact (b), the fault impact signal added with Gaussian white noise (c) and the envelope spectrum thereof (d) are shown, and it can be found that the signal is interfered by strong noise and the characteristics of the fault are completely submerged.
[0053] S1, sound signals are collected by using an array microphone sensor and a related data collection device, as shown in Figure 5 , which are original signals collected by each sensor in the embodiment;
[0054] S2, a high expected beam response map and a low expected beam response map are obtained for a target azimuth angle, the corresponding target azimuth angle being 60°, and then a high-power response beamformer and a low-power response beamformer are calculated by using a second-order cone optimization method; in the embodiment, the calculated high-power response beamformer and low-power response beamformer are as shown in Figure 6 .
[0055] S3, when the target signal, the interference signal and the target signal+interference signal exist respectively, the signals collected by the array microphone sensor are input into the high-power response beamformer and the low-power response beamformer, the collected signals of each sensor are processed into a high-power output signal and a low-power output signal by vector operation of the high-power beamformer and the low-power beamformer, and then the signals output by the high-power beamformer and the low-power beamformer are subjected to difference processing, and the result is as shown in Figure 7 , it can be seen that the interference signal can be significantly weakened and the target signal can be reserved by performing difference processing on the sound signals output by the high-power response beamformer and the low-power response beamformer respectively.
[0056] S4, the difference value output by the difference power response beam is subjected to CEEMDAN decomposition and noise reduction processing, and the processing result is as shown in Figure 8 , the component with the largest kurtosis is selected, and the component screening result is as shown in Table 1. The MOMEDA is used to strengthen the fault impact component, and finally the processed signal is subjected to envelope demodulation analysis, as shown in Figure 9 , thereby realizing rolling bearing fault diagnosis based on acoustic signals.
[0057] Table 1 Component screening result
[0058] MODE 1 2 3 4 5 kurtosis 2.126 7.02 2.87 3.16 3.05 MODE 6 7 8 9 … kurtosis 3.11 2.75 2.92 2.71 …
[0059] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions recorded in the above embodiments can be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.
Claims
1. A method for diagnosing acoustic faults in rolling bearings with controllable noise power, characterized in that, Includes the following steps: S1. Acquire the sound signal of the rolling bearing using a microphone array sensor; S2. A high-power response beamformer and a low-power response beamformer are established based on the second-order cone programming method. S3. The sound array signal is processed by a high-power response beamformer and a low-power response beamformer. The outputs of the two beamformers are differentially processed to obtain the desired target acoustic signal with controllable noise. S4. Based on the desired target acoustic signal obtained in step S3, extract fault features and determine the fault type; In step S2, the specific method for establishing the high-power response beamformer and the low-power response beamformer based on the second-order cone programming method is as follows: (1) Initialize the sidelobe beam level and set a high-expectation beam response map and a low-expectation beam response map for the target azimuth angle; the responsivity of the high-expectation beam response map at the target azimuth angle is higher than that of the low-expectation beam response map at the target azimuth angle. (2) Based on the lowest sidelobe beam design method, calculate the high-power response beamformer and the low-power beamformer; the calculation formula is: ； in, w H represents the weighted vector of the beamformer, and H represents the conjugate transpose. Indicates azimuth. Indicates the direction of the main lobe. Indicates the beam width of the sidelobe level. This represents the microphone array manifold vector. This represents the beam response diagram.
2. The method for diagnosing acoustic faults in rolling bearings with controllable noise power according to claim 1, characterized in that, Step S4 also includes a step of filtering the desired target acoustic signal.
3. The method for diagnosing acoustic faults in rolling bearings with controllable noise power according to claim 2, characterized in that, In step S4, the desired target acoustic signal is filtered using the adaptive noise complete set empirical mode decomposition method.
4. The method for diagnosing acoustic faults in rolling bearings with controllable noise power according to claim 2, characterized in that, In step S4, after filtering the desired target acoustic signal, the method further includes a step of feature enhancement of the desired target acoustic signal.
5. The method for diagnosing acoustic faults in rolling bearings with controllable noise power according to claim 4, characterized in that, In step S4, multi-point optimal minimum entropy deconvolution is used to enhance signal features.
6. The method for diagnosing acoustic faults in rolling bearings with controllable noise power according to claim 1, characterized in that, In step S4, the desired target acoustic signal is analyzed by envelope demodulation to extract fault features and determine the fault type.
7. The method for diagnosing acoustic faults in rolling bearings with controllable noise power according to claim 1, characterized in that, Step S1 further includes analyzing the directional information of the target bearing sound source and noise source, and determining the optimal values for the microphone linear array spacing and the sound source distance.