Bearing fault detection method and system based on improved filtering algorithm
Through the improved wavelet threshold noise reduction algorithm and S-G filter, the bearing vibration signals of the rotating equipment of the sewage treatment plant are pre-processed. Combined with time-frequency analysis and envelope demodulation, the problem of serious noise-affected bearing vibration signals is solved, and high-accurate bearing fault detection is achieved.
Patent Information
- Application Number
- CN202411944255.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-27
- Publication Date
- 2025-04-25
AI Technical Summary
The bearing vibration signal of the rotating equipment of the sewage treatment plant is severely affected by noise, which makes it difficult to extract fault characteristics and cannot accurately determine the bearing failure.
The original vibration data is preprocessed with an improved wavelet threshold noise reduction algorithm and an S-G filter, and the fault detection of bearing vibration signals is performed through time-frequency analysis and envelope demodulation.
Effectively suppress noise, extract bearing failure characteristics, improve the accuracy of fault detection, and adapt to bearing failure detection needs under complex units.
Smart Images
Figure CN119880429A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of engineering industry technology, and more specifically, to a bearing fault detection method and system based on an improved filtering algorithm. Background Art
[0002] As a key component in mechanical equipment or mechanical transmission system, bearings are widely used in rotating equipment such as motors, gearboxes, dehydrators, submersible pumps, flow thrusters and blowers in sewage treatment plants. During the operation of the equipment, bearings will have outer ring, inner ring, rolling element and other faults over time, which will seriously affect the stability of equipment operation and cause production safety accidents. Therefore, it is of great significance to effectively detect the vibration signals of equipment operation.
[0003] The rotating equipment in sewage treatment plants, especially underwater equipment, has complex service conditions and is affected by changes in operating conditions, environmental noise, resonance noise between different equipment, etc., which causes the effective bearing vibration signal to be submerged in the above-mentioned invalid noise, making it difficult to extract the effective bearing fault features and unable to accurately determine the bearing fault.
[0004] In order to solve the above problems, it is necessary to perform data preprocessing on the original vibration signal of the equipment, suppress or remove the invalid components in the original vibration signal through noise reduction, time-frequency processing, etc., and extract valid feature data. Summary of the invention
[0005] In order to solve the above technical problems, the present invention proposes a bearing fault detection method and system based on an improved filtering algorithm, the purpose of which is to pre-process the original vibration data by improving the wavelet threshold denoising algorithm and the SG filter, and to perform fault detection on the bearing vibration signal in the motor under the complex unit through time-frequency analysis and envelope demodulation.
[0006] A first aspect of the present invention provides a bearing fault detection method based on an improved filtering algorithm, comprising the following steps:
[0007] Collecting a vibration signal of a target bearing, and performing noise reduction processing on the vibration signal using an improved wavelet threshold filter to obtain a noise-reduced vibration signal;
[0008] Performing time-frequency analysis on the noise reduction vibration signal to obtain a time-domain waveform diagram and a spectrum feature, judging whether the target bearing is faulty and determining a preliminary fault diagnosis result based on the time-domain waveform diagram and the spectrum feature;
[0009] Determine the data source of envelope demodulation according to the speed data of the shaft where the target bearing is located, and use the noise reduction vibration signal when the speed data does not exceed the preset speed threshold, otherwise use the original vibration signal;
[0010] The envelope signal is obtained by Hilbite-Huang transform, and the SG filter is set based on the rotation frequency of the shaft where the target bearing is located and the bearing fault frequency range. The envelope noise reduction signal is obtained by filling and filtering the envelope signal data.
[0011] The spectrum of the envelope denoised signal is obtained, and the bearing fault features in the envelope spectrum are extracted. The bearing fault type is further determined based on the bearing fault features in the envelope spectrum, and the final fault diagnosis result of the target bearing is obtained.
[0012] In this scheme, the wavelet threshold denoising algorithm is improved by constraining the wavelet threshold filter threshold function between the hard threshold and the soft threshold, so that the threshold function is continuous in the wavelet domain and the reconstruction error is reduced. The improved threshold function is expressed as:
[0013]
[0014] in is the wavelet coefficient, w is the original wavelet coefficient, λ is the threshold, and sign(·) is the sign function;
[0015] The vibration signal X of the target bearing acc The improved wavelet threshold denoising algorithm is used to remove some high-frequency noise in the original vibration signal to obtain the denoised vibration signal X nr .
[0016] In this solution, whether the target bearing is faulty is determined based on the time domain waveform and spectrum characteristics and a preliminary fault diagnosis result is determined, specifically:
[0017] Get the noise reduction vibration signal X nr A time domain waveform diagram is obtained, and the shape and impact of the waveform are read according to the time domain waveform diagram to obtain the time domain performance characteristics, and the preset judgment standard is determined by integrating the time domain performance of the fault sample through the fault instance, and the similarity between the time domain performance characteristics and the preset judgment standard is calculated, and a preliminary judgment is made on whether the target bearing has a fault according to the similarity;
[0018] Obtaining the noise-reduced vibration signal X through fast Fourier transform nr The frequency spectrum of the bearing is used to extract the bearing fault characteristics in the spectrum, including the outer ring frequency, inner ring frequency, rolling element frequency, cage frequency and the multiples of each frequency. The proportion of the bearing fault characteristics in the frequency spectrum is compared with the frequency spectrum of the healthy bearing vibration signal to determine the preliminary fault diagnosis result of the target bearing.
[0019] In this scheme, after determining the data source of the envelope demodulation, the envelope of the data source signal is extracted through the Hilbite-Huang transform, and the envelope signal X is obtained by extracting the absolute value of the envelope. es .
[0020] In this scheme, the bearing fault frequency and its multiple frequency range are determined by using the calculation formula of the rolling element frequency based on the basic mechanical parameters of the target bearing, and the design parameters of the SG filter are determined according to the rotation frequency of the shaft where the target bearing is located and the shaft fault frequency range. The design parameters of the SG filter include the window length and the number of polynomial terms.
[0021] When the rotation frequency of the shaft where the target bearing is located is lower than 1Hz, the window length is selected as 99. When the rotation frequency is higher than 1Hz, the window length decreases accordingly. The change of the window length win_size is expressed as:
[0022]
[0023] where f i is the rotation frequency of the shaft of the equipment where the target bearing is located. The calculated window length needs to be rounded down. If win_size is an even number, the window length is win_size-1;
[0024] Based on the highest fault frequency f of the target bearing max Determine the number of terms K of the polynomial of the SG filter, and its judgment condition is expressed as:
[0025]
[0026] The SG filter is set using adaptively selected design parameters.
[0027] In this scheme, the envelope signal X es The edge data in is filled with data to obtain the envelope signal X after data filling optimization tc , X tc Input to the SG filter for filtering to obtain the envelope noise reduction signal X ns .
[0028] In this scheme, the envelope demodulation signal X is obtained by fast Fourier transform ns The spectrum of the bearing is obtained, and the bearing fault characteristics in the envelope spectrum are extracted. By analyzing the bearing fault characteristics in the envelope spectrum, the proportion of the bearing fault characteristics in the envelope spectrum in the spectrum is compared with the spectrum of the healthy bearing vibration signal to further determine the bearing fault type.
[0029] The bearing fault type is combined with the initial fault diagnosis result to determine the final fault diagnosis result of the target bearing, and the bearing fault type is obtained.
[0030] The second aspect of the present invention provides a bearing fault detection system based on an improved filtering algorithm, the system comprising: a vibration signal acquisition unit, a wavelet threshold filtering unit, a time-frequency analysis unit, an envelope demodulation unit, an SG filter unit, and a comprehensive analysis unit;
[0031] The vibration signal acquisition unit is responsible for acquiring the vibration signal of the target bearing;
[0032] The wavelet threshold filtering unit improves the wavelet threshold denoising algorithm, and uses the improved wavelet threshold filtering to perform denoising on the collected vibration signal to obtain a denoised vibration signal;
[0033] The time-frequency analysis unit performs time-frequency analysis on the noise reduction vibration signal to obtain a time domain waveform diagram and spectrum characteristics, judge whether the target bearing is faulty and determine a preliminary fault diagnosis result;
[0034] The envelope demodulation unit uses Hilt-Huang transform to perform envelope demodulation on the noise reduction vibration signal or the original vibration signal corresponding to the target bearing to obtain an envelope signal;
[0035] The SG filter unit adaptively sets the SG filter parameters based on the rotation frequency of the shaft where the target bearing is located and the bearing fault frequency range, and obtains the envelope noise reduction signal by filling and filtering the data of the envelope signal;
[0036] The comprehensive analysis unit obtains the spectrum of the envelope noise reduction signal, extracts the bearing fault characteristics in the envelope spectrum, further determines the bearing fault type, and obtains the final fault diagnosis result of the target bearing in combination with the initial fault diagnosis result analysis.
[0037] Compared with the prior art, the beneficial effects of this solution are:
[0038] Aiming at the problem that the bearing vibration signal of equipment in complex units is seriously affected by noise, the threshold function of wavelet threshold denoising is improved and the SG filter with adaptive design parameters and boundary optimization is designed to adapt to the characteristics of the signal, eliminate the influence of noise and ensure the accuracy of detection.
[0039] The bearing fault and fault type can be comprehensively judged through time domain waveform, spectrum characteristics, envelope demodulation and theoretical experience, which is close to the actual needs of online or on-site monitoring of industrial equipment. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.
[0041] Figure 1 A flow chart of a bearing fault detection method based on an improved filtering algorithm of the present invention is shown;
[0042] Figure 2A time domain waveform diagram of a noise reduction vibration signal of an embodiment is shown;
[0043] Figure 3 shows the spectrum of the noise-reduced vibration signal of the embodiment;
[0044] Figure 4 The spectrum of the target bearing noise reduction envelope signal in the embodiment is shown;
[0045] Figure 5 A block diagram of a bearing fault detection system based on an improved filtering algorithm of the present invention is shown. DETAILED DESCRIPTION
[0046] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0047] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.
[0048] Figure 1 A flow chart of a bearing fault detection method based on an improved filtering algorithm of the present invention is shown.
[0049] like Figure 1 As shown, in the first embodiment of the present invention, a bearing fault detection method based on an improved filtering algorithm is provided, comprising:
[0050] Step 1: collecting the vibration signal of the target bearing, and performing noise reduction processing on the vibration signal using an improved wavelet threshold filter to obtain a noise-reduced vibration signal;
[0051] It should be noted that the default threshold functions of wavelet threshold denoising are soft threshold and hard threshold, but these two threshold functions are not continuous and will produce certain errors in signal reconstruction. The wavelet threshold denoising algorithm is improved by constraining the wavelet threshold filter threshold function between the hard threshold and the soft threshold. By rewriting the reconstruction threshold function, the wavelet coefficients are constrained between the hard threshold and the soft threshold, making the threshold function continuous in the wavelet domain and reducing the reconstruction error. The improved threshold function is expressed as:
[0052]
[0053] in is the wavelet coefficient, w is the original wavelet coefficient, λ is the threshold, the threshold used in this scheme is sqtwolog, and sign(·) is the sign function.
[0054] The vibration signal X of the target bearing acc =[x1,x2,...,x i ,...,x n ], where x i 、x n is the i-th sampling point and the n-th sampling point of the vibration signal, n is the total number of sampling points, and the improved wavelet threshold denoising algorithm is used to remove some high-frequency noise in the original vibration signal. In this embodiment, in the improved wavelet threshold denoising, the number of wavelet decomposition layers is 3, the wavelet base is sym4 wavelet, mln is used as the decomposition scale, and the improved threshold function is used. Reconstruct the original signal to obtain the noise-reduced vibration signal X nr , to reduce the impact of noise.
[0055] Step 2: Perform time-frequency analysis on the noise reduction vibration signal to obtain a time domain waveform and spectrum characteristics, and determine whether the target bearing is faulty and determine a preliminary fault diagnosis result based on the time domain waveform and spectrum characteristics;
[0056] It should be noted that obtaining the noise reduction vibration signal X nr The time domain waveform of Figure 2 As shown. The time domain performance characteristics are obtained by reading the shape and impact of the waveform according to the time domain waveform diagram, and the preset judgment standard is determined by integrating the time domain performance of the fault sample through the fault instance. The similarity between the time domain performance characteristics and the preset judgment standard is calculated, and the target bearing is preliminarily judged whether it is faulty according to the similarity. For example, when an obvious sudden or periodic impact is detected in the time domain waveform of the signal, and the duration of the impact is long, it can be considered that the bearing has an obvious fault. In addition, when a more obvious waveform change occurs, such as a sinusoidal change trend, it can also be determined that the bearing may have a fault; and when a weak sudden or periodic impact is detected in the time domain waveform of the signal, and the duration of the impact is short, it can be considered that the bearing may have a fault.
[0057] In this embodiment, if Figure 3 As shown, the time domain performance characteristics are constructed by obtaining the periodic impacts and impact intervals that appear in the time domain waveform diagram of the target bearing noise reduction vibration signal. The similarity calculation is performed using the time series performance characteristics to know the rolling element failure frequency close to the bearing. Therefore, it can be preliminarily determined that the equipment has a fault and subsequent fault detection is required to confirm the fault.
[0058] Obtaining the noise-reduced vibration signal X through fast Fourier transform nr The spectrum, such as Figure 3The bearing fault characteristics in the spectrum are extracted, including the outer ring frequency (BPFO), inner ring frequency (BPFI), rolling element frequency (BSF), cage frequency (FSF) and the multiples of each frequency. The specific calculation formulas of BPFO, BPFI, BSF and FSF are as follows:
[0059]
[0060] Where d is the diameter of the bearing roller, D is the bearing median diameter, α is the contact angle, Z is the number of rollers, and f i For frequency conversion.
[0061] In this example, the speed of the shaft where the target bearing is located is 3000 rpm / min, the frequency is 50 Hz, and the fault frequency of the bearing is shown in Table 1 below:
[0062] Table 1 Bearing failure frequency of the drive motor in the blower of a sewage treatment plant
[0063]
[0064] The proportion of the bearing fault feature in the frequency spectrum is compared with the frequency spectrum of the healthy bearing vibration signal to determine the preliminary fault diagnosis result of the target bearing. In this example, in the horizontal direction, the inner ring fault frequency and the rolling element fault frequency of the target bearing are very obvious and have a high amplitude; therefore, it can be preliminarily determined that the bearing of the equipment has a fault, and the fault may be caused by the inner ring or rolling element of the bearing.
[0065] Step 3: Determine the data source of the envelope demodulation according to the speed data of the shaft where the target bearing is located, and use the noise reduction vibration signal when the speed data does not exceed the preset speed threshold, otherwise use the original vibration signal;
[0066] Preferably, the speed threshold is set to 60 rpm / min. When the speed of the shaft where the bearing is located is lower than the speed threshold, the device type is a low-speed device, and the noise-reduced vibration signal X after wavelet threshold noise reduction is taken. nr In this example, since the speed of the shaft where the target bearing is located is 3000rpm / min, the original vibration signal X is used in subsequent signal processing. acc .
[0067] Step 4: Obtain the envelope signal through the Hilbite-Huang transform, set the SG filter based on the rotation frequency of the shaft where the target bearing is located and the bearing fault frequency range, and obtain the envelope noise reduction signal by filling and filtering the envelope signal data;
[0068] It should be noted that the fault signal of the bearing in a complex unit often has low impact energy and high noise ratio. It is impossible to accurately judge the fault and the fault type only by the time domain waveform and spectrum. After determining the data source of the envelope demodulation, the envelope of the data source signal is extracted through the Hilbit-Huang transform, and the envelope signal X is obtained by extracting the absolute value of the envelope. es The envelope signal is less affected by noise than the vibration signal.
[0069] Based on the basic mechanical parameters of the target bearing, the calculation formula of the rolling element frequency is used to determine the frequency range of the bearing fault frequency and its multiples. The design parameters of the SG filter are determined according to the rotation frequency of the shaft where the target bearing is located and the shaft fault frequency range. The design parameters of the SG filter include the window length and the number of polynomial terms. The process of adaptively determining the design parameters of the SG filter is as follows:
[0070] When the rotation frequency of the shaft where the target bearing is located is lower than 1Hz (6rpm / s), the window length is selected as 99. When the rotation frequency is higher than 1Hz, the window length decreases accordingly. The change in window length win_size is expressed as:
[0071]
[0072] where f i is the rotation frequency of the shaft of the equipment where the target bearing is located. The calculated window length needs to be rounded down. If win_size is an even number, the window length is win_size-1;
[0073] Based on the highest fault frequency f of the target bearing max Determine the number of terms K of the polynomial of the SG filter, and its judgment condition is expressed as:
[0074]
[0075] The SG filter is set using adaptively selected design parameters. The noise reduction capability and ease of use of the SG filter are improved through adaptive parameter selection and data filling.
[0076] For the envelope signal X es The edge data in is filled with data to obtain the envelope signal X after data filling optimization tc , which avoids the situation where the filter fails to process effectively when some signal data is located at the boundary. The specific filling formula is:
[0077]
[0078] Where a, b, c, x, y, z all represent the envelope signal X es The edge data in , c, b, a, z, y, x are the filling data.
[0079] In this embodiment, the bearing fault frequency range required by SG is between [50Hz, 730Hz]. According to the adaptive parameter selection, the SG filter is determined to be a 3-point 5-order low-pass filter. es The edge data of the optimized envelope signal X is filled with at least 10 edge data. tc Input to the SG filter for filtering to obtain the envelope noise reduction signal X ns .
[0080] Step 5: Obtain the spectrum of the envelope denoised signal and extract the bearing fault features in the envelope spectrum. Based on the bearing fault features in the envelope spectrum, further determine the bearing fault type and obtain the final fault diagnosis result of the target bearing.
[0081] It should be noted that the envelope demodulation signal X is obtained by fast Fourier transform ns The spectrum of the noise reduction envelope signal is obtained by extracting the bearing fault features in the envelope spectrum. By analyzing the bearing fault features in the envelope spectrum and comparing the proportion of the bearing fault features in the envelope spectrum with the spectrum of the healthy bearing vibration signal, the bearing fault type is further determined. In this example, the spectrum of the noise reduction envelope signal is as follows: Figure 5 As shown, the cage frequency of the target bearing is excited, and the rolling element fault frequency is also very obviously excited. Therefore, it is further determined that the bearing in the drive motor of the faulty fan is faulty. Based on the conclusion that the fault is caused by damage to the inner ring or rolling element of the bearing in the preliminary fault diagnosis results of the target bearing, a comprehensive analysis is conducted to determine that the target bearing has failed or the equipment bearing is insufficiently lubricated, and it is necessary to shut down in time for troubleshooting. Through actual inspection, it is determined that the lubricating oil in the equipment has deteriorated and the bearing is worn, verifying the final fault diagnosis result.
[0082] Figure 5 A block diagram of a bearing fault detection system based on an improved filtering algorithm of the present invention is shown.
[0083] The second embodiment of the present invention provides a bearing fault detection system based on an improved filtering algorithm, the system comprising: a vibration signal acquisition unit, a wavelet threshold filtering unit, a time-frequency analysis unit, an envelope demodulation unit, an SG filter unit, and a comprehensive analysis unit;
[0084] The vibration signal acquisition unit is responsible for acquiring the vibration signal of the target bearing;
[0085] The wavelet threshold filtering unit improves the wavelet threshold denoising algorithm, and uses the improved wavelet threshold filtering to perform denoising on the collected vibration signal to obtain a denoised vibration signal;
[0086] The time-frequency analysis unit performs time-frequency analysis on the noise reduction vibration signal to obtain a time domain waveform diagram and spectrum characteristics, judge whether the target bearing is faulty and determine a preliminary fault diagnosis result;
[0087] The envelope demodulation unit uses Hilt-Huang transform to perform envelope demodulation on the noise reduction vibration signal or the original vibration signal corresponding to the target bearing to obtain an envelope signal;
[0088] The SG filter unit adaptively sets the SG filter parameters based on the rotation frequency of the shaft where the target bearing is located and the bearing fault frequency range, and obtains the envelope noise reduction signal by filling and filtering the data of the envelope signal;
[0089] The comprehensive analysis unit obtains the spectrum of the envelope noise reduction signal, extracts the bearing fault characteristics in the envelope spectrum, further determines the bearing fault type, and obtains the final fault diagnosis result of the target bearing in combination with the initial fault diagnosis result analysis.
[0090] The third embodiment of the present invention provides a computer-readable storage medium, which includes a bearing fault detection method program based on an improved filtering algorithm. When the bearing fault detection method program based on the improved filtering algorithm is executed by a processor, the steps of the bearing fault detection method based on the improved filtering algorithm are implemented.
[0091] In the several embodiments provided in the present application, it should be understood that the disclosed method can be implemented in other ways. The embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0092] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.
[0093] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.
[0094] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A bearing fault detection method based on an improved filtering algorithm, characterized in that: The following steps are involved: Collecting a vibration signal of a target bearing, and performing noise reduction processing on the vibration signal using an improved wavelet threshold filter to obtain a noise-reduced vibration signal; Performing time-frequency analysis on the noise reduction vibration signal to obtain a time-domain waveform diagram and a spectrum feature, judging whether the target bearing is faulty and determining a preliminary fault diagnosis result based on the time-domain waveform diagram and the spectrum feature; Determine the data source of envelope demodulation according to the speed data of the shaft where the target bearing is located, and use the noise reduction vibration signal when the speed data does not exceed the preset speed threshold, otherwise use the original vibration signal; The envelope signal is obtained by Hilbite-Huang transform, and the SG filter is set based on the rotation frequency of the shaft where the target bearing is located and the bearing fault frequency range. The envelope noise reduction signal is obtained by filling and filtering the envelope signal data. The spectrum of the envelope denoised signal is obtained, and the bearing fault features in the envelope spectrum are extracted. The bearing fault type is further determined based on the bearing fault features in the envelope spectrum, and the final fault diagnosis result of the target bearing is obtained.
2. The bearing fault detection method based on the improved filtering algorithm according to claim 1 is characterized in that: The wavelet threshold filtering threshold function is constrained between the hard threshold and the soft threshold to improve the wavelet threshold denoising algorithm, so that the threshold function is continuous in the wavelet domain and the reconstruction error is reduced. The improved threshold function is expressed as: in is the wavelet coefficient, w is the original wavelet coefficient, λ is the threshold, and sign(·) is the sign function; The vibration signal X of the target bearing acc The improved wavelet threshold denoising algorithm is used to remove some high-frequency noise in the original vibration signal to obtain the denoised vibration signal X nr .
3. The bearing fault detection method based on the improved filtering algorithm according to claim 1 is characterized in that: According to the time domain waveform and spectrum characteristics, it is judged whether the target bearing is faulty and the preliminary fault diagnosis result is determined, specifically: Get the noise reduction vibration signal X nr A time domain waveform diagram is obtained, and the shape and impact of the waveform are read according to the time domain waveform diagram to obtain the time domain performance characteristics, and the preset judgment standard is determined by integrating the time domain performance of the fault sample through the fault instance, and the similarity between the time domain performance characteristics and the preset judgment standard is calculated, and a preliminary judgment is made on whether the target bearing has a fault according to the similarity; Obtaining the noise-reduced vibration signal X through fast Fourier transform nr The frequency spectrum of the bearing is used to extract the bearing fault characteristics in the spectrum, including the outer ring frequency, inner ring frequency, rolling element frequency, cage frequency and the multiples of each frequency. The proportion of the bearing fault characteristics in the frequency spectrum is compared with the frequency spectrum of the healthy bearing vibration signal to determine the preliminary fault diagnosis result of the target bearing.
4. The bearing fault detection method based on the improved filtering algorithm according to claim 1 is characterized in that: After determining the data source of the envelope demodulation, the envelope of the data source signal is extracted through the Hilbite-Huang transform, and the envelope signal X is obtained by extracting the absolute value of the envelope. es .
5. The bearing fault detection method based on the improved filtering algorithm according to claim 1 is characterized in that: Based on the basic mechanical parameters of the target bearing, the calculation formula of the rolling element frequency is used to determine the frequency range of the bearing fault frequency and its multiples, and the design parameters of the SG filter are determined according to the rotation frequency of the shaft where the target bearing is located and the shaft fault frequency range. The design parameters of the SG filter include the window length and the number of polynomial terms; When the rotation frequency of the shaft where the target bearing is located is lower than 1Hz, the window length is selected as 99. When the rotation frequency is higher than 1Hz, the window length decreases accordingly. The change of the window length win_size is expressed as: where f i is the rotation frequency of the shaft of the equipment where the target bearing is located. The calculated window length needs to be rounded down. If win_size is an even number, the window length is win_size-1; Based on the highest fault frequency f of the target bearing max Determine the number of terms K of the polynomial of the SG filter, and its judgment condition is expressed as: The SG filter is set using adaptively selected design parameters.
6. The bearing fault detection method based on the improved filtering algorithm according to claim 1 is characterized in that: For the envelope signal X es The edge data in is filled with data to obtain the envelope signal X after data filling optimization tc , X tc Input to the SG filter for filtering to obtain the envelope noise reduction signal X ns .
7. The bearing fault detection method based on the improved filtering algorithm according to claim 1 is characterized in that: The envelope demodulation signal X is obtained by fast Fourier transform ns The spectrum of the bearing is obtained, and the bearing fault characteristics in the envelope spectrum are extracted. By analyzing the bearing fault characteristics in the envelope spectrum, the proportion of the bearing fault characteristics in the envelope spectrum in the spectrum is compared with the spectrum of the healthy bearing vibration signal to further determine the bearing fault type. The bearing fault type is combined with the initial fault diagnosis result to determine the final fault diagnosis result of the target bearing, and the bearing fault type is obtained.
8. A bearing fault detection system based on an improved filtering algorithm, characterized in that: Implementing the bearing fault detection method based on the improved filtering algorithm as described in any one of claims 1 to 7, the system comprises: a vibration signal acquisition unit, a wavelet threshold filtering unit, a time-frequency analysis unit, an envelope demodulation unit, an SG filter unit, and a comprehensive analysis unit; The vibration signal acquisition unit is responsible for acquiring the vibration signal of the target bearing; The wavelet threshold filtering unit improves the wavelet threshold denoising algorithm, and uses the improved wavelet threshold filtering to perform denoising on the collected vibration signal to obtain a denoised vibration signal; The time-frequency analysis unit performs time-frequency analysis on the noise reduction vibration signal to obtain a time domain waveform diagram and spectrum characteristics, judge whether the target bearing is faulty and determine a preliminary fault diagnosis result; The envelope demodulation unit uses Hilt-Huang transform to perform envelope demodulation on the noise reduction vibration signal or the original vibration signal corresponding to the target bearing to obtain an envelope signal; The SG filter unit adaptively sets the SG filter design parameters based on the rotation frequency of the shaft where the target bearing is located and the bearing fault frequency range, and obtains the envelope noise reduction signal by filling and filtering the data of the envelope signal; The comprehensive analysis unit obtains the spectrum of the envelope noise reduction signal, extracts the bearing fault characteristics in the envelope spectrum, further determines the bearing fault type, and obtains the final fault diagnosis result of the target bearing in combination with the initial fault diagnosis result analysis.