Acoustic emission-based variable pitch bearing on-line state monitoring system

Abstract

The invention discloses a variable pitch bearing online state monitoring system based on acoustic emission, and relates to the technical field of bearing monitoring. The system comprises a first acoustic emission receiving unit, a second acoustic emission receiving unit, an elastic wave emission unit and an integrated host. The elastic wave transmitting unit is used for transmitting an inspection signal; the first acoustic emission receiving unit and the second acoustic emission receiving unit are both used for receiving inspection signals and elastic waves generated by bearing movement; and the integrated host receives signals from the acoustic emission acquisition card, and analyzes and processes elastic force waves through variable pitch bearing on-line state monitoring software taking a virtual instrument technology as a platform. The first acoustic emission receiving unit and the second acoustic emission receiving unit are used for collecting the elastic wave generated by the motion of the variable pitch bearing, the CNN fault recognition model is used for recognizing the elastic wave to judge the damage condition of the bearing, abnormal working of the variable pitch bearing is found in time, the service life of the variable pitch bearing is prolonged, and the cost is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of bearing monitoring, and in particular relates to an online state monitoring system for a variable pitch bearing based on acoustic emission. Background Art

[0002] According to the characteristics of blades falling during the operation of wind turbines, the main reasons are as follows: first, the pitch bearing cracks, causing the inner and outer rings of the bearing to separate; second, the bolts connecting the pitch bearing and the hub break; third, the bolts connecting the pitch bearing and the blade break. For example, during the service life of a certain megawatt-class model, due to the lack of full consideration of stress concentration and corrosion prevention in a humid environment, the bolt holes connecting the blades and the pitch bearings rusted after a period of operation, which made the stress more concentrated, and then aggravated the rust of the bolt holes, and finally led to the cracking of the pitch bearing. The cracking process starts from the inside of the bolt hole to the bolt hole cracking through. After the lubricating oil is squeezed out, the pitch bearing continues to receive alternating loads, and finally cracks are formed on the pitch bearing. In addition, the bolts connecting the pitch bearing and the hub and the bolts connecting the pitch bearing and the blade often break during the operation of existing megawatt-class units. In the case of minor cases, they are discovered in time and some or all bolts are replaced in time. In the case of serious cases, the blades fall directly. To prevent cracking of the pitch bearing during operation, some units have added reinforcement rings to the pitch bearings and an online crack alarm system consisting of enameled wire buried in the reinforcement rings. However, there are still false alarms or late alarms, and there is still a risk of cracks expanding and blades falling.

[0003] There are many ways to lubricate pitch bearings, such as oil injection lubrication, oil-gas lubrication, etc. Their common feature is to form a lubricating oil film of a certain thickness between the rolling element and the inner and outer rings of the bearing, reduce friction and heat generation in the contact area, and ensure the good operation of the bearing. If poor lubrication leads to difficulty in forming the oil film or the oil film breaks, the friction and wear inside the bearing will increase, and the heat generation will increase sharply, greatly affecting the service performance and service life of the bearing.

[0004] In addition, the bearings of wind turbines will make abnormal noises when they are initially damaged. If these acoustic signals can be captured in time, when the pitch bearings show early signs of fatigue, single-point lubrication and the use of anti-wear lubricants can effectively delay their failure under extreme working conditions. Summary of the invention

[0005] The purpose of the present invention is to provide an online condition monitoring system for a variable pitch bearing based on acoustic emission. The elastic waves generated by the movement of the variable pitch bearing are collected by a first acoustic emission receiving unit and a second acoustic emission receiving unit, and the elastic waves are identified by a CNN fault recognition model to determine the damage of the bearing, thereby solving the problem that the initial damage of the existing variable pitch bearing cannot be monitored, resulting in a short service life and high cost of the variable pitch bearing.

[0006] In order to solve the above technical problems, the present invention is achieved through the following technical solutions: The present invention is an online condition monitoring system for a variable pitch bearing based on acoustic emission, comprising a plurality of first acoustic emission receiving units evenly distributed on the outer ring of the bearing, a plurality of second acoustic emission receiving units evenly distributed on the inner ring of the bearing, an elastic wave emission unit at least arranged on one bearing roller, and an integrated host; The elastic wave transmitting unit is used to transmit the inspection signal; The first acoustic emission receiving unit and the second acoustic emission receiving unit are both used to receive the inspection signal and the elastic wave generated by the bearing movement; the first acoustic emission receiving unit and the second acoustic emission receiving unit are both electrically connected to a signal conditioning circuit; the signal conditioning circuit is used to obtain the inspection signal received by the first acoustic emission receiving unit and the second acoustic emission receiving unit, and transmit it to the acoustic emission acquisition card for acquisition and then send it to the integrated host through the PCI interface; The integrated host is equipped with online status monitoring software for variable pitch bearings and an alarm display unit circuit.

[0007] As a preferred technical solution, the signal conditioning circuit includes a preamplifier and a bandpass filter; the preamplifier is electrically connected to the bandpass filter and is used to amplify the inspection signal; the bandpass filter is used to allow waves in a specific frequency band to pass through and simultaneously shield waves in other frequency bands; the preamplifier is a device used to test the working status of the acoustic emission detection system itself or to indicate the connection; its main functions are: (1) providing impedance matching between the high-impedance sensor and the low-impedance transmission bearing to prevent signal attenuation; (2) improving the signal-to-noise ratio related to bearing noise by amplifying weak input signals; (3) reducing the common-mode electrical noise introduced by the sensor and its bearing through differential amplification; (4) providing a frequency filter; the bandpass filter sets the frequency band allowed to pass between 10kHz-5000kHz.

[0008] As a preferred technical solution, the workflow of the online condition monitoring software for the variable pitch bearing is as follows: Step S1: Collect a large number of elastic waves generated by the movement of the pitch bearing and train a CNN fault recognition model; Step S2: Separating and processing the collected elastic waves; Step S3: Using wavelet packet energy distribution vector and Mel logarithm to extract elastic wave signal features, and inputting them into CNN fault recognition model; Step S4: The CNN fault recognition model identifies elastic wave features and identifies faults.

[0009] As a preferred technical solution, in step S1, the training process of the CNN fault recognition model is as follows: Step S11: extracting waveform characteristics of elastic waves generated by the movement of the pitch bearing; Step S12: Design a CNN fault recognition model structure, including a convolutional layer, a pooling layer, and a fully connected layer; Step S13: using the elastic wave waveform features as a training set to train the CNN model; Step S14: Use the validation set to monitor the performance of the model during the training process, and adjust the model parameters according to the validation results; Step S15: Use the test set to evaluate the trained model.

[0010] As a preferred technical solution, in step S11, the elastic wave waveform feature extraction process is as follows: Step S111: Use the Hamming window to smooth the waveform frequency. The specific formula is: ; Where M is the window length, and n is an integer from 0 to (M-1); Step S112: Substitute the logarithm of the power spectrum into the Mel frequency cepstrum coefficient. The specific formula is as follows: ; In the formula, is the actual frequency; Step S113: Convert the time domain data into frequency domain data and perform Fourier transform on each frame signal. The specific formula is: , where represents the i-th frame, represents frequency domain data, Represents time domain data; Step S114: square the frequency domain data to obtain the capacity of the spectrum line. The specific formula is: ; Step S115: Compare the frequency response of the Mel filter in the frequency domain Energy spectrum of each frame Multiplication and addition are equivalent, and the specific formula is: ; Step S116: taking the logarithm of the calculation result of the Mel filter and performing discrete cosine transform to obtain a feature vector. The specific formula is as follows: ; Where i is the i-th frame, n is the order of MFCC, m is the m-th filter, and M is the number of filters. is the Mel filter energy.

[0011] As a preferred technical solution, in step S12, multiple convolution kernel filters convolve the original input image to extract features, the convolution layer is pooled by the pooling layer to reduce the extracted feature size, the learned features are mapped to the sample label space by the fully connected layer, and finally the results are output through the softmax classifier and the fully connected layer.

[0012] As a preferred technical solution, the first acoustic emission receiving unit and the second acoustic emission receiving unit both include an upper electrode layer, a piezoelectric layer, a lower electrode layer and a substrate layer, a cavity is formed in the substrate layer, an adjustable elastic component is provided in the cavity, and both ends of the elastic component are respectively pressed against the inner wall of the cavity.

[0013] As a preferred technical solution, the elastic component includes a magnet and a winding coupled to the magnet, and the winding is electrically connected to the control chip.

[0014] The present invention has the following beneficial effects: The present invention collects elastic waves generated by the movement of the variable pitch bearing through the first acoustic emission receiving unit and the second acoustic emission receiving unit, and uses the CNN fault recognition model to identify the elastic waves to determine the damage of the bearing, so as to timely discover the abnormal operation of the variable pitch bearing, increase the service life of the variable pitch bearing and reduce the cost.

[0015] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0017] Figure 1 It is a structural schematic diagram of an online condition monitoring system for a pitch bearing based on acoustic emission of the present invention; Figure 2 This is a schematic diagram of the structure of the pitch bearing. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0020] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1-2 It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0021] Embodiment 1 See also Figure 1 As shown, the working process of the online condition monitoring software 701 for the variable pitch bearing is as follows: Step S1: Collect a large number of elastic waves generated by the movement of the pitch bearing and train a CNN fault recognition model; Step S2: Separating and processing the collected elastic waves; Step S3: Using wavelet packet energy distribution vector and Mel logarithm to extract elastic wave signal features, and inputting them into CNN fault recognition model; Step S4: The CNN fault recognition model identifies elastic wave features and identifies faults.

[0022] In step S1, the training process of the CNN fault recognition model is as follows: Step S11: extracting waveform characteristics of elastic waves generated by the movement of the pitch bearing; Step S12: Design a CNN fault recognition model structure, including a convolutional layer, a pooling layer, and a fully connected layer; Step S13: using the elastic wave waveform features as a training set to train the CNN model; Step S14: Use the validation set to monitor the performance of the model during the training process, and adjust the model parameters according to the validation results; Step S15: Use the test set to evaluate the trained model.

[0023] In step S11, the elastic wave waveform feature extraction process is as follows: Step S111: Use the Hamming window to smooth the waveform frequency. The specific formula is: ; Where M is the window length, and n is an integer from 0 to (M-1); Step S112: Substitute the logarithm of the power spectrum into the Mel frequency cepstrum coefficient. The specific formula is as follows: ; In the formula, is the actual frequency; Step S113: Convert the time domain data into frequency domain data and perform Fourier transform on each frame signal. The specific formula is: , where represents the i-th frame, represents frequency domain data, Represents time domain data; Step S114: square the frequency domain data to obtain the capacity of the spectrum line. The specific formula is: ; Step S115: Compare the frequency response of the Mel filter in the frequency domain Energy spectrum of each frame Multiplication and addition are equivalent, and the specific formula is: ; Step S116: taking the logarithm of the calculation result of the Mel filter and performing discrete cosine transform to obtain a feature vector. The specific formula is as follows: ; Where i is the i-th frame, n is the order of MFCC, m is the m-th filter, and M is the number of filters. is the Mel filter energy.

[0024] Embodiment 2 See also Figure 2 As shown, the present invention is an online condition monitoring system for a pitch bearing based on acoustic emission, which can be used to execute the method content of Example 1 of the present invention, including a plurality of first acoustic emission receiving units 1 evenly distributed on the outer ring 101 of the bearing, and a plurality of second acoustic emission receiving units 2 evenly distributed on the inner ring 201 of the bearing, an elastic wave emission unit 3 at least arranged on one bearing roller 301, and an integrated host 7; The elastic wave transmitting unit 3 is used to transmit the inspection signal; The first acoustic emission receiving unit 1 and the second acoustic emission receiving unit 2 are both used to receive the inspection signal and the elastic wave generated by the bearing movement; the first acoustic emission receiving unit 1 and the second acoustic emission receiving unit 2 are both electrically connected to the signal conditioning circuit 4; the signal conditioning circuit 4 is used to obtain the inspection signal received by the first acoustic emission receiving unit 1 and the second acoustic emission receiving unit 2, and transmit it to the acoustic emission acquisition card 5 for acquisition and then send it to the integrated host 7 through the PCI interface 6; The integrated host 7 has built-in variable pitch bearing online status monitoring software 701 and an alarm display unit circuit 702 .

[0025] The first acoustic emission receiving unit 1 and the second acoustic emission receiving unit 2 both include an upper electrode layer, a piezoelectric layer, a lower electrode layer and a substrate layer. A cavity is formed in the substrate layer. An adjustable elastic component is provided in the cavity. Both ends of the elastic component are respectively pressed against the inner wall of the cavity.

[0026] The elastic component includes a magnet and a winding coupled with the magnet, and the winding is electrically connected to the control chip.

[0027] It is worth noting that in the above system embodiment, the various units included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0028] In addition, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiments can be completed by instructing related hardware through a program, and the corresponding program can be stored in a computer-readable storage medium.

[0029] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An online condition monitoring system for a variable pitch bearing based on acoustic emission, comprising a plurality of first acoustic emission receiving units (1) evenly distributed on a bearing outer ring (101), a plurality of second acoustic emission receiving units (2) evenly distributed on a bearing inner ring (201), an elastic wave transmitting unit (3) arranged on at least one bearing roller (301), and an integrated host (7), characterized in that: The elastic wave transmitting unit (3) is used to transmit an inspection signal; The first acoustic emission receiving unit (1) and the second acoustic emission receiving unit (2) are both used to receive inspection signals and elastic waves generated by bearing movement; the first acoustic emission receiving unit (1) and the second acoustic emission receiving unit (2) are both electrically connected to a signal conditioning circuit (4); the signal conditioning circuit (4) is used to obtain the inspection signals received by the first acoustic emission receiving unit (1) and the second acoustic emission receiving unit (2), and transmit them to the acoustic emission acquisition card (5) for acquisition and then transmit them to the integrated host (7) via the PCI interface (6); The integrated host (7) has built-in variable pitch bearing online status monitoring software (701) and an alarm display unit circuit (702); the integrated host (7) receives a signal from an acoustic emission acquisition card (5), and analyzes and processes elastic waves through the variable pitch bearing online status monitoring software (701) based on virtual instrument technology.

2. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 1 is characterized in that: The signal conditioning circuit (4) comprises a preamplifier (401), a bandpass filter (402) and a digital operator (403); the preamplifier (401) is electrically connected to the bandpass filter (402) and is used to amplify the inspection signal; the bandpass filter (402) is used to allow waves in a specific frequency band to pass through while shielding waves in other frequency bands; the digital operator (403) is electrically connected to the bandpass filter (402) and is used to convert the electrical signal into a digital signal.

3. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 1 is characterized in that: The working process of the online condition monitoring software (701) for the variable pitch bearing is as follows: Step S1: Collect a large number of elastic waves generated by the movement of the pitch bearing and train a CNN fault recognition model; Step S2: Separating and processing the collected elastic waves; Step S3: Using wavelet packet energy distribution vector and Mel logarithm to extract elastic wave signal features, and inputting them into CNN fault recognition model; Step S4: The CNN fault recognition model identifies elastic wave features and identifies faults.

4. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 3 is characterized in that: In step S1, the training process of the CNN fault recognition model is as follows: Step S11: extracting waveform characteristics of elastic waves generated by the movement of the pitch bearing; Step S12: Design a CNN fault recognition model structure, including a convolutional layer, a pooling layer, and a fully connected layer; Step S13: using the elastic wave waveform features as a training set to train the CNN model; Step S14: Use the validation set to monitor the performance of the model during the training process, and adjust the model parameters according to the validation results; Step S15: Use the test set to evaluate the trained model.

5. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 4 is characterized in that: In step S11, the elastic wave waveform feature extraction process is as follows: Step S111: Use the Hamming window to smooth the waveform frequency. The specific formula is: ； Where M is the window length, and n is an integer from 0 to (M-1); Step S112: Substitute the logarithm of the power spectrum into the Mel frequency cepstrum coefficient. The specific formula is as follows: ； In the formula, is the actual frequency; Step S113: Convert the time domain data into frequency domain data and perform Fourier transform on each frame signal. The specific formula is: , where represents the i-th frame, represents frequency domain data, Represents time domain data; Step S114: square the frequency domain data to obtain the capacity of the spectrum line. The specific formula is: ; Step S115: Compare the frequency response of the Mel filter in the frequency domain Energy spectrum of each frame Multiplication and addition are equivalent, and the specific formula is: ； Step S116: taking the logarithm of the calculation result of the Mel filter and performing discrete cosine transform to obtain a feature vector. The specific formula is as follows: ； Where i is the i-th frame, n is the order of MFCC, m is the m-th filter, and M is the number of filters. is the Mel filter energy.

6. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 4 is characterized in that: In step S12, multiple convolution kernel filters convolve the original input image to extract features, the convolution layer is pooled by the pooling layer to reduce the extracted feature size, the learned features are mapped to the label space of the sample by the fully connected layer, and finally the results are output through the softmax classifier and the fully connected layer.

7. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 1 is characterized in that: The first acoustic emission receiving unit (1) and the second acoustic emission receiving unit (2) both comprise an upper electrode layer, a piezoelectric layer, a lower electrode layer and a substrate layer, a cavity is formed in the substrate layer, an adjustable elastic component is provided in the cavity, and two ends of the elastic component are respectively pressed against the inner wall of the cavity.

8. The online condition monitoring system for pitch bearing based on acoustic emission according to claim 7 is characterized in that: The elastic component includes a magnet and a winding coupled with the magnet, and the winding is electrically connected to the control chip.

Images