Bearing fault diagnosis method and device based on MVMD noise reduction and fusing impact pulse and stress wave

Abstract

The embodiment of the invention provides a bearing fault diagnosis method and device based on MVMD noise reduction and fusing impact pulses and stress waves and electronic equipment, and belongs to the technical field of fault diagnos.The method comprises the steps that a vibration sensor is adopted to collect vibration signal data when an inner ring and an outer ring of a bearing break down; decomposing the collected fault vibration signal data by using a multi-variable variational mode decomposition (MVMD) mode; calculating a correlation coefficient between each IMF and an original fault vibration signal; analyzing the reconstructed signal by using an impact pulse algorithm, calculating an impact characteristic value, and distinguishing an inner ring fault from an outer ring fault of the bearing by using a K-nearest neighbor classification algorithm according to the impact characteristic value and stress wave characteristic parameters to realize automatic identification of fault types; and drawing an envelope spectrogram of the fault signal, and determining a specific part of the fault by analyzing a specific frequency band and a peak value in the envelope spectrogram. According to the scheme, the bearing fault can be judged timely and accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a bearing fault diagnosis method, device and electronic equipment based on MVMD noise reduction and fusion of impact pulses and stress waves. Background Art

[0002] Bearing fault diagnosis refers to monitoring and analyzing the operating status of bearings through various technologies and methods to identify and evaluate whether there is a bearing fault and the type and severity of the fault. Bearings are widely used rotating support components in mechanical equipment, and their performance directly affects the operating efficiency and safety of the entire mechanical system. Therefore, timely and accurate bearing fault diagnosis is of great significance for preventing equipment failures, extending service life and reducing maintenance costs.

[0003] Although the existing bearing fault diagnosis technology has made significant progress, there are still some shortcomings and challenges in practical application. The following are some common shortcomings:

[0004] 1. Data collection and sensor limitations

[0005] Sensor location and quantity: The installation location and quantity of sensors have a significant impact on data quality and diagnostic results. Improper sensor arrangement may result in incomplete or distorted data.

[0006] Sensor performance: The accuracy, frequency response range, and stability of the sensor directly affect the reliability of the data. Low-cost or low-quality sensors may introduce noise and errors.

[0007] Environmental factors: Sensors may be affected by external factors such as environmental noise and electromagnetic interference, resulting in data distortion.

[0008] 2. Data processing and analysis complexity

[0009] Signal processing difficulty: Bearing fault signals usually contain multiple frequency components and noise, requiring complex signal processing techniques (such as filtering, decomposition, feature extraction, etc.) to separate and extract fault features.

[0010] Feature selection and extraction: Selecting appropriate feature parameters (such as peak factor, pulse energy, stress wave energy, etc.) is a challenge, and different fault types may require different feature combinations.

[0011] Computational resources: Advanced signal processing and data analysis methods (such as deep learning, multivariate variational mode decomposition, etc.) require a lot of computing resources and may not be suitable for real-time monitoring systems.

[0012] 3. Fault identification and classification accuracy

[0013] False positives and false negatives: Existing fault diagnosis methods may suffer from the problems of false positives and false negatives, especially in early fault detection.

[0014] Identification of multiple fault types: There are various types of bearing faults, including inner ring faults, outer ring faults, ball faults, etc. Existing methods may have limitations in identifying multiple fault types.

[0015] Nonlinear failure: The development process of bearing failure is often nonlinear, and existing linear models may not be able to accurately describe and predict such failures.

[0016] 4. Real-time performance and response speed

[0017] Real-time monitoring: In industrial environments, real-time monitoring and rapid response are key requirements. Existing methods may require long processing time and complex calculations, which are difficult to meet the needs of real-time monitoring.

[0018] Data transmission delay: In distributed monitoring systems, data transmission delay may affect real-time performance and accuracy.

[0019] 5. Maintenance and cost

[0020] System maintenance: The long-term operation of sensors and data acquisition systems requires regular maintenance and calibration, which increases maintenance costs and workload.

[0021] Hardware cost: High-performance sensors and data acquisition equipment are expensive, which may limit their popularity and use in certain application scenarios.

[0022] 6. Data Management and Storage

[0023] Large amount of data: Long-term monitoring will generate a large amount of data, which requires effective data management and storage solutions, otherwise it will lead to data loss or management difficulties.

[0024] Data security: In the industrial Internet and cloud computing environment, data security and privacy protection have become an important issue.

[0025] The above problems become technical problems that need to be solved. Summary of the invention

[0026] In view of this, an embodiment of the present invention provides a bearing fault diagnosis method, device and electronic equipment based on MVMD noise reduction and fusion of impact pulses and stress waves, which at least partially solve the problems existing in the prior art.

[0027] In a first aspect, an embodiment of the present invention provides a bearing fault diagnosis method based on MVMD noise reduction and fusion of shock pulses and stress waves, comprising:

[0028] S1: using a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, wherein the vibration signal data includes fault information;

[0029] S2: Decompose the collected fault vibration signal data using multivariate variational mode decomposition (MVMD) to obtain multiple basic modal components (IMFs), each of which represents the signal characteristics in a different frequency range;

[0030] S3: Calculate the correlation coefficient between each IMF and the original fault vibration signal, and screen out the IMF with a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal;

[0031] S4: Select the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold, and add these IMFs to construct a reconstructed signal to highlight the fault characteristics;

[0032] S5: applying the shock pulse algorithm to analyze the reconstructed signal and calculating the shock characteristic value, which is used to describe the shock intensity and frequency of the fault location;

[0033] S6: further process the reconstructed signal using a stress wave algorithm to calculate stress wave characteristic parameters, wherein the stress wave characteristic parameters include stress wave energy, pulse energy, and maximum peak height, and describe the severity and type of the fault based on the stress wave characteristic parameters;

[0034] S7: Using the K nearest neighbor classification algorithm, the inner ring fault of the bearing is distinguished from the outer ring fault according to the impact characteristic value and stress wave characteristic parameters, and the fault type is automatically identified;

[0035] S8: Draw the envelope spectrum of the fault signal, and confirm the specific location of the fault by analyzing the specific frequency bands and peaks in the envelope spectrum.

[0036] According to a specific implementation of an embodiment of the present invention, the vibration sensor is used to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, and the vibration signal data contains fault information, including:

[0037] Use vibration sensors mounted on the mechanical structure near the bearings;

[0038] After the vibration signal is collected by the data acquisition system, the collected vibration signal is amplified, filtered and digitized to obtain the final collected signal.

[0039] According to a specific implementation of an embodiment of the present invention, the multivariate variational mode decomposition (MVMD) method is used to decompose the collected fault vibration signal data to obtain multiple basic modal components (IMFs), including:

[0040] Import the collected multi-channel vibration signal data into the data processing software and store the data in the form of time series, with each time point corresponding to one or more human vibration signal values;

[0041] Determine the MVMD decomposition layer number K and the regularization parameter α, determine the number of IMFs after decomposition based on the decomposition layer number K, and control the smoothness and sparsity of each modal component in the decomposition process based on the regularization parameter α;

[0042] The objective function of MVMD is constructed to minimize the difference between the original signal and the sum of the decomposed modal components. At the same time, a regularization term is introduced to control the smoothness and sparsity of the modal components. The objective function is expressed as:

[0043]

[0044] Among them, x k is the original vibration signal of the kth channel, u i is the i-th modal component, α is the regularization parameter;

[0045] Use numerical optimization method to solve the above objective function and get each modal component u i .

[0046] According to a specific implementation of an embodiment of the present invention, the calculation of the correlation coefficient between each IMF and the original fault vibration signal, and screening out the IMF having a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal, includes:

[0047] K IMFs are obtained through multivariate variational mode decomposition, denoted as IMF1, IMF2, …, IMF K , prepare the original fault vibration signal data, denoted as x(t);

[0048] For each IMF i , calculate IMF i The correlation coefficient with the original signal x(t):

[0049]

[0050] Where N is the length of the signal, is the mean of the original signal x(t), It is the IMF i The mean of (t);

[0051] Set the correlation coefficient threshold ρ, for each IMF i ,if It is considered that the IMF is highly correlated with the original signal and the IMF is retained. i .

[0052] According to a specific implementation of an embodiment of the present invention, selecting the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold, and adding these IMFs to construct the reconstructed signal includes:

[0053] Add the screened highly correlated IMFs to construct the reconstructed signal y(t):

[0054]

[0055] Where I is the index set of the screened highly correlated IMFs.

[0056] According to a specific implementation of an embodiment of the present invention, the application of the shock pulse algorithm to analyze the reconstructed signal and calculate the shock characteristic value includes:

[0057] Calculate the peak factor CF included in the shock characteristic value:

[0058]

[0059] Calculate the peak energy PE of the signal:

[0060] PE=(max(|y(t)|)) 2

[0061] Calculate the impulse value II of the signal:

[0062]

[0063] According to a specific implementation of an embodiment of the present invention, the stress wave algorithm is used to further process the reconstructed signal to calculate the stress wave energy, pulse energy and maximum peak height, including:

[0064] Use a high-pass filter H(t) to extract the high-frequency components in the signal;

[0065] Calculate the stress wave energy SE:

[0066]

[0067] Calculate the pulse energy PE:

[0068] PE=(max(|y(t)·H(t)|)) 2

[0069] Calculate the maximum peak height MP:

[0070] MP = max(|y(t)·H(t)|).

[0071] According to a specific implementation of an embodiment of the present invention, the use of a K nearest neighbor classification algorithm to distinguish a bearing inner ring fault from an outer ring fault according to an impact characteristic value and a stress wave characteristic parameter includes:

[0072] Collect a training data set with known labels to ensure that the data set contains samples of inner circle faults and outer circle faults;

[0073] The impact characteristic value and stress wave characteristic parameters are extracted for each sample, and the features are processed using Zscore standardization;

[0074] Select the K value through cross-validation, use the training data set to train the KNN classifier, and use the test data set to verify the performance of the classifier;

[0075] Perform feature extraction and standardization on the new samples, and use the trained KNN classifier for classification.

[0076] In a second aspect, an embodiment of the present invention provides a bearing fault diagnosis device based on MVMD noise reduction and fusion of shock pulses and stress waves, comprising:

[0077] A collection module, using a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing fail, wherein the vibration signal data includes failure information;

[0078] The decomposition module uses the multivariate variational mode decomposition (MVMD) method to decompose the collected fault vibration signal data to obtain multiple basic modal components (IMFs), each of which represents the signal characteristics within a different frequency range.

[0079] The comparison module calculates the correlation coefficient between each IMF and the original fault vibration signal, and screens out the IMF with a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal;

[0080] The reconstruction module selects the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold and adds these IMFs to construct a reconstructed signal to highlight the fault characteristics;

[0081] The analysis module applies the shock pulse algorithm to analyze the reconstructed signal and calculates the shock characteristic value, which is used to describe the shock intensity and frequency of the fault location;

[0082] A calculation module, which further processes the reconstructed signal using a stress wave algorithm to calculate stress wave characteristic parameters, wherein the stress wave characteristic parameters include stress wave energy, pulse energy, and maximum peak height, and describes the severity and type of the fault based on the stress wave characteristic parameters;

[0083] The differentiation module uses the K nearest neighbor classification algorithm to distinguish the inner ring fault from the outer ring fault of the bearing according to the impact characteristic value and stress wave characteristic parameters, thereby realizing automatic identification of the fault type;

[0084] The drawing module draws the envelope spectrum of the fault signal and confirms the specific location of the fault by analyzing the specific frequency bands and peaks in the envelope spectrum.

[0085] In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0086] at least one processor; and,

[0087] a memory communicatively connected to the at least one processor; wherein,

[0088] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the bearing fault diagnosis method based on MVMD noise reduction, fusion of impact pulses and stress waves in any of the aforementioned first aspects or any implementation of the first aspect.

[0089] In a fourth aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable the computer to execute the bearing fault diagnosis method based on MVMD noise reduction, fusion of shock pulses and stress waves in the aforementioned first aspect or any implementation method of the first aspect.

[0090] In the fifth aspect, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the bearing fault diagnosis method based on MVMD noise reduction, fusion of impact pulses and stress waves in the aforementioned first aspect or any implementation of the first aspect.

[0091] The bearing fault diagnosis scheme based on MVMD noise reduction, fusion of impact pulses and stress waves in the embodiment of the present invention includes: using a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, and the vibration signal data contains fault information; using multivariate variational mode decomposition MVMD to decompose the collected fault vibration signal data to obtain multiple basic modal components IMF, each IMF represents the signal characteristics in different frequency ranges; calculating the correlation coefficient between each IMF and the original fault vibration signal, and by comparing the correlation between each IMF and the original signal, screening out IMFs with a high correlation with the fault vibration signal; selecting basic modal components IMFs with correlation coefficients greater than a preset threshold, and combining these IMFs The reconstructed signal is constructed by adding them together to highlight the fault characteristics; the reconstructed signal is analyzed by applying the shock pulse algorithm to calculate the shock characteristic value, which is used to describe the shock intensity and frequency of the fault location; the reconstructed signal is further processed by the stress wave algorithm to calculate the stress wave characteristic parameters, which include stress wave energy, pulse energy and maximum peak height, and the severity and type of the fault are described based on the stress wave characteristic parameters; the K nearest neighbor classification algorithm is used to distinguish the inner ring fault from the outer ring fault of the bearing according to the shock characteristic value and the stress wave characteristic parameter, so as to realize automatic identification of the fault type; the envelope spectrum of the fault signal is drawn, and the specific location of the fault is confirmed by analyzing the specific frequency bands and peak values ​​in the envelope spectrum.

[0092] The solution of this application has the following beneficial effects:

[0093] 1. Enhanced noise reduction capability: The adaptive time-frequency characteristics of the MVMD algorithm make the filtering effect not limited by a fixed frequency band, which is especially suitable for processing nonlinear and non-stationary vibration signals, effectively improving the signal-to-noise ratio and reducing the probability of false triggering.

[0094] 2. Improve fault identification accuracy: By extracting rich vibration signal features through shock pulse and stress wave algorithms, it can more comprehensively and accurately reflect the equipment status and improve the accuracy and reliability of fault identification.

[0095] 3. Realize early fault warning: Using shock pulse and stress wave technology, it is possible to capture early fault signs, shorten fault detection time, and improve the timeliness and efficiency of equipment maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments 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 based on these drawings without paying creative work.

[0097] Figure 1 A flow chart of a bearing fault diagnosis method based on MVMD noise reduction and fusion of shock pulses and stress waves provided in an embodiment of the present invention;

[0098] Figure 2 A schematic flow chart of another bearing fault diagnosis method based on MVMD noise reduction and fusion of shock pulses and stress waves provided in an embodiment of the present invention;

[0099] Figure 3 A schematic diagram of the structure of a bearing fault diagnosis device based on MVMD noise reduction and fusion of shock pulses and stress waves provided in an embodiment of the present invention;

[0100] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0101] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0102] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0103] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0104] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0105] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0106] The embodiment of the present disclosure provides a bearing fault diagnosis method based on MVMD noise reduction, fusion of shock pulses and stress waves. The bearing fault diagnosis method based on MVMD noise reduction, fusion of shock pulses and stress waves provided in this embodiment can be executed by a computing device, which can be implemented as software, or as a combination of software and hardware, and the computing device can be integrated in a server, terminal device, etc.

[0107] See also Figure 1 The embodiment of the present disclosure provides a bearing fault diagnosis method based on MVMD noise reduction and fusion of shock pulses and stress waves, including:

[0108] S1: Using a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing fail, the vibration signal data contains failure information.

[0109] A vibration sensor is used to collect vibration signal data when the inner and outer rings of the bearing fail. The vibration signal data contains fault information. The specific implementation method is as follows:

[0110] 1. Choose the right vibration sensor:

[0111] Type: Choose a high-sensitivity, wide-bandwidth accelerometer that can capture the high- and low-frequency vibration signals generated by the bearing during operation.

[0112] Performance parameters: Make sure the sensor has a good frequency response range (usually 0-10kHz), high resolution, low noise, and high dynamic range.

[0113] Environmental adaptability: Select sensors that can work normally in harsh environments such as high temperature, humidity, and electromagnetic interference.

[0114] 2. Sensor installation location:

[0115] Inner ring fault monitoring: Install the sensor in the radial position of the bearing seat, close to the inner ring. Inner ring faults usually produce high-frequency vibrations, and the sensor should be as close to the inner ring as possible to capture these high-frequency signals.

[0116] Outer ring fault monitoring: Install the sensor in the axial position of the bearing seat, close to the outer ring. Outer ring faults usually produce medium-frequency vibrations, and the sensor should be as close to the outer ring as possible to capture these medium-frequency signals.

[0117] Multi-point monitoring: To comprehensively monitor the condition of the bearing, sensors can be installed in multiple locations, such as radial, axial and lateral locations of the bearing seat.

[0118] 3.Sensor fixing method:

[0119] Magnetic fixation: Use the magnetic base to fix the sensor on a metal surface, suitable for temporary or mobile monitoring.

[0120] Bolt fixing: Use bolts to fix the sensor to the bearing seat, which is suitable for long-term monitoring and ensures good contact between the sensor and the object being measured.

[0121] Adhesive Fixing: Use high-strength adhesive to fix the sensor on non-metallic surfaces for special occasions.

[0122] 4. Data acquisition system:

[0123] Data acquisition card: Using a high-precision data acquisition card can convert the analog signals collected by the sensor into digital signals.

[0124] Sampling rate: Set an appropriate sampling rate, usually several times the fault characteristic frequency. For example, if the fault characteristic frequency of the bearing is 1kHz, the sampling rate should be set to at least 5kHz.

[0125] Filtering and amplification: Appropriate filters and amplifiers are added to the data acquisition system to remove noise and enhance the signal.

[0126] 5. Data Transmission and Storage:

[0127] Wired transmission: Use cables to connect the sensor to the data acquisition system, which is suitable for short-distance monitoring.

[0128] Wireless transmission: Use wireless sensor networks (WSN) to transmit data to a central data acquisition system, suitable for long-distance monitoring.

[0129] Data storage: The collected vibration signal data is stored locally or in a cloud server to ensure data security and integrity.

[0130] 6. Data preprocessing:

[0131] Denoising: Use a digital filter (such as a low-pass filter, high-pass filter, or band-pass filter) to remove noise from a signal.

[0132] Normalization: Normalize the data to eliminate the dimensional differences between different sensors.

[0133] Time domain and frequency domain analysis: Perform time domain and frequency domain analysis on vibration signals to extract time domain features (such as mean, variance, peak factor, etc.) and frequency domain features (such as spectrum, power spectral density, etc.).

[0134] Specific steps may include:

[0135] 1. Choose the right vibration sensor:

[0136] Select the appropriate acceleration sensor according to the working environment and fault characteristics of the bearing.

[0137] 2. Sensor installation location:

[0138] Install sensors at radial locations on the bearing housing to monitor inner race faults.

[0139] A sensor is installed in the axial position of the bearing seat to monitor outer ring faults.

[0140] If necessary, install sensors at multiple locations to fully monitor the bearing condition.

[0141] 3.Sensor fixing method:

[0142] Use magnets, bolts or adhesive to fix the sensor to the bearing housing, ensuring good contact between the sensor and the object being measured.

[0143] 4. Data acquisition system:

[0144] Use a high-precision data acquisition card to convert the analog signals collected by the sensor into digital signals.

[0145] Set an appropriate sampling rate to ensure that the characteristic frequency of the fault can be captured.

[0146] Add appropriate filters and amplifiers to the data acquisition system to remove noise and enhance the signal.

[0147] 5. Data Transmission and Storage:

[0148] Transmit data to a central data acquisition system using wired or wireless methods.

[0149] The collected vibration signal data is stored locally or in a cloud server to ensure data security and integrity.

[0150] 6. Data preprocessing:

[0151] Use digital filters to remove noise from the signal.

[0152] The data is normalized to eliminate the dimensional differences between different sensors.

[0153] The vibration signal is analyzed in time domain and frequency domain to extract characteristic parameters.

[0154] Through the above steps, the vibration sensor can be effectively used to collect vibration signal data when the inner ring and outer ring of the bearing fail. The vibration signal data contains fault information and provides reliable data support for subsequent fault diagnosis and analysis.

[0155] S2: The collected fault vibration signal data is decomposed using the multivariate variational mode decomposition (MVMD) method to obtain multiple basic modal components (IMFs), each of which represents the signal characteristics within a different frequency range.

[0156] The collected fault vibration signal data is decomposed using the multivariate variational mode decomposition (MVMD) method to obtain multiple fundamental mode components (IMFs). Each IMF represents the signal characteristics in a different frequency range. The specific implementation method is as follows:

[0157] 1. Prepare data:

[0158] Ensure that multi-channel fault vibration signal data has been collected through the vibration sensor. Assume that we have N channels of vibration signal data, and the data of each channel is recorded as x1(t), x2(t),…, x N (t).

[0159] 2. Select MVMD parameters:

[0160] Decomposition level K: Determines the number of IMFs to be decomposed. The decomposition level K should be selected based on the characteristics of the signal and experimental experience. Generally, the more decomposition levels there are, the finer the decomposed IMFs will be, but the computational complexity will also increase.

[0161] Regularization parameter α: controls the smoothness and sparsity of each modal component during the decomposition process. The choice of regularization parameter α can be determined experimentally or empirically.

[0162] 3. Construct the objective function:

[0163] The objective function of MVMD aims to minimize the difference between the original signal and the sum of the decomposed modal components, while introducing a regularization term to control the smoothness and sparsity of the modal components. The objective function can be expressed as:

[0164]

[0165] Among them, x k is the original vibration signal of the kth channel, u i is the i-th modal component, α is the regularization parameter

[0166] 4. Solve optimization problems:

[0167] The objective function is solved using numerical optimization methods (such as the alternating direction multiplier method ADMM) to obtain the modal components. The ADMM method iteratively updates each variable and gradually approaches the optimal solution.

[0168] Through the above steps, the multivariate variational mode decomposition (MVMD) method can be effectively used to decompose the collected fault vibration signal data to obtain multiple fundamental mode components (IMFs). Each IMF represents the signal characteristics in a different frequency range, providing a basis for subsequent fault feature extraction and diagnosis.

[0169] S3: Calculate the correlation coefficient between each IMF and the original fault vibration signal, and screen out the IMF with a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal.

[0170] The correlation coefficient between each IMF and the original fault vibration signal is calculated, and the IMF with a high correlation with the fault vibration signal is screened out by comparing the correlation between each IMF and the original signal. The specific implementation method is as follows:

[0171] K IMFs are obtained through multivariate variational mode decomposition, denoted as IMF1, IMF2, …, IMF K , prepare the original fault vibration signal data, denoted as x(t);

[0172] For each IMF i , calculate IMF i The correlation coefficient with the original signal x(t):

[0173]

[0174] Where N is the length of the signal, is the mean of the original signal x(t), It is the IMF i The mean of (t);

[0175] Set the correlation coefficient threshold ρ, for each IMF i ,if It is considered that the IMF is highly correlated with the original signal and the IMF is retained. i .

[0176] The screened highly correlated IMFs are saved as independent time series data files for further analysis and processing.

[0177] Through the above steps, the correlation coefficient between each IMF and the original fault vibration signal can be effectively calculated, and the IMF with a high correlation with the fault vibration signal can be screened out, providing high-quality data support for subsequent fault feature extraction and diagnosis.

[0178] S4: Select the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold, and add these IMFs to construct a reconstructed signal to highlight the fault characteristics.

[0179] The highly correlated IMFs selected are added together to construct a reconstructed signal to highlight the fault characteristics. The specific implementation method is as follows:

[0180] It is ensured that multiple fundamental mode components (IMFs) have been obtained through the multivariate variational mode decomposition (MVMD) method, and the correlation coefficient between each IMF and the original fault vibration signal is calculated.

[0181] The IMFs with correlation coefficients greater than a preset threshold are screened out, and these IMFs are considered to be highly correlated with the fault vibration signal.

[0182] Assume that the highly correlated IMFs that have been screened out are IMF_1, IMF_2, \ldots, IMF_m, where m is the number of highly correlated IMFs.

[0183] Add the highly correlated IMFs that have been screened out to construct the reconstructed signal y(t). The calculation formula for the reconstructed signal y(t) is as follows:

[0184] Add the screened highly correlated IMFs to construct the reconstructed signal y(t):

[0185]

[0186] Where I is the index set of the screened highly correlated IMFs.

[0187] Through the above steps, the screened highly correlated IMFs can be effectively added to construct the reconstructed signal, thereby highlighting the fault characteristics and providing high-quality data support for subsequent fault diagnosis.

[0188] S5: Apply the shock pulse algorithm to analyze the reconstructed signal and calculate the shock characteristic value, which is used to describe the shock intensity and frequency of the fault location.

[0189] The peak factor CF included in the shock characteristic value can be calculated:

[0190]

[0191] Calculate the peak energy PE of the signal:

[0192] PE=(max(|y(t)|)) 2

[0193] Calculate the impulse value II of the signal:

[0194]

[0195] S6: further process the reconstructed signal using a stress wave algorithm to calculate stress wave characteristic parameters, which include stress wave energy, pulse energy, and maximum peak height, and describe the severity and type of the fault based on the stress wave characteristic parameters.

[0196] A high-pass filter H(t) can be used to extract the high-frequency components in the signal;

[0197] Calculate the stress wave energy SE:

[0198]

[0199] Calculate the pulse energy PE:

[0200] PE=(max(|y(t)·H(t)|)) 2

[0201] Calculate the maximum peak height MP:

[0202] MP = max(|y(t)·H(t)|).

[0203] S7: Using the K nearest neighbor classification algorithm, the inner ring fault of the bearing is distinguished from the outer ring fault according to the impact characteristic value and stress wave characteristic parameters, and the fault type is automatically identified.

[0204] Specifically, a training data set with known labels can be collected to ensure that the data set contains samples of inner ring faults and outer ring faults; the impact characteristic value and stress wave characteristic parameter are extracted for each sample, and the features are processed using Zscore normalization; the K value is selected through cross-validation, the KNN classifier is trained using the training data set, and the performance of the classifier is verified using the test data set; feature extraction and normalization are performed on new samples, and the trained KNN classifier is used for classification.

[0205] S8: Draw the envelope spectrum of the fault signal, and confirm the specific location of the fault by analyzing the specific frequency bands and peaks in the envelope spectrum.

[0206] Draw the envelope spectrum of the fault signal and confirm the specific location of the fault by observing the specific frequency bands and peaks in the envelope spectrum. The specific implementation method is as follows:

[0207] Perform envelope demodulation on the reconstructed signal y(t) to extract the envelope of the signal. Envelope demodulation usually includes the following steps:

[0208] Perform Hilbert transform on the reconstructed signal y(t) to obtain the analytical signal z(t):

[0209]

[0210] in, is the Hilbert transform of y(t).

[0211] The amplitude of the analytical signal z(t) is the signal envelope e(t):

[0212]

[0213] Perform Fourier transform on the extracted envelope e(t) to obtain the spectrum E(f) of the envelope:

[0214]

[0215] in, is the Fourier transform of e(t).

[0216] Use a drawing tool (such as Matplotlib) to draw an envelope spectrum to display the frequency distribution and peak value of the envelope spectrum E(f).

[0217] Analyze the specific frequency bands and peaks in the envelope spectrum to confirm the specific location of the fault. Common fault characteristic frequencies include:

[0218] Fundamental frequency: The frequency at which a bearing rotates.

[0219] Inner race fault characteristic frequency: Inner race fault usually has a peak at an integer multiple of the fundamental frequency.

[0220] Outer race fault characteristic frequency: Outer race fault usually has a peak at a non-integer multiple of the fundamental frequency.

[0221] Ball bearing failure characteristic frequency: Ball bearing failure usually has peaks at the higher harmonics of the fundamental frequency.

[0222] According to a specific implementation of an embodiment of the present invention, the vibration sensor is used to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, and the vibration signal data contains fault information, including:

[0223] Use vibration sensors mounted on the mechanical structure near the bearings;

[0224] After the vibration signal is collected by the data acquisition system, the collected vibration signal is amplified, filtered and digitized to obtain the final collected signal.

[0225] According to a specific implementation of an embodiment of the present invention, the multivariate variational mode decomposition (MVMD) method is used to decompose the collected fault vibration signal data to obtain multiple basic modal components (IMFs), including:

[0226] Import the collected multi-channel vibration signal data into the data processing software and store the data in the form of time series, with each time point corresponding to one or more human vibration signal values;

[0227] Determine the MVMD decomposition layer number K and the regularization parameter α, determine the number of IMFs after decomposition based on the decomposition layer number K, and control the smoothness and sparsity of each modal component in the decomposition process based on the regularization parameter α;

[0228] The objective function of MVMD is constructed to minimize the difference between the original signal and the sum of the decomposed modal components. At the same time, a regularization term is introduced to control the smoothness and sparsity of the modal components. The objective function is expressed as:

[0229]

[0230] Among them, x k is the original vibration signal of the kth channel, u i is the i-th modal component, α is the regularization parameter;

[0231] Use numerical optimization method to solve the above objective function and get each modal component u i .

[0232] According to a specific implementation of an embodiment of the present invention, the calculation of the correlation coefficient between each IMF and the original fault vibration signal, and screening out the IMF having a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal, includes:

[0233] K IMFs are obtained through multivariate variational mode decomposition, denoted as IMF1, IMF2, …, IMF K , prepare the original fault vibration signal data, denoted as x(t);

[0234] For each IMF i , calculate IMF i The correlation coefficient with the original signal x(t):

[0235]

[0236] Where N is the length of the signal, is the mean of the original signal x(t), It is the IMF i The mean of (t);

[0237] Set the correlation coefficient threshold ρ, for each IMF i ,if It is considered that the IMF is highly correlated with the original signal and the IMF is retained. i .

[0238] According to a specific implementation of an embodiment of the present invention, selecting the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold, and adding these IMFs to construct the reconstructed signal includes:

[0239] Add the screened highly correlated IMFs to construct the reconstructed signal y(t):

[0240]

[0241] Where I is the index set of the screened highly correlated IMFs.

[0242] According to a specific implementation of an embodiment of the present invention, the application of the shock pulse algorithm to analyze the reconstructed signal and calculate the shock characteristic value includes:

[0243] Calculate the peak factor CF included in the shock characteristic value:

[0244]

[0245] Calculate the peak energy PE of the signal:

[0246] PE=(max(|y(t)|)) 2

[0247] Calculate the impulse value II of the signal:

[0248]

[0249] According to a specific implementation of an embodiment of the present invention, the stress wave algorithm is used to further process the reconstructed signal to calculate the stress wave energy, pulse energy and maximum peak height, including:

[0250] Use a high-pass filter H(t) to extract the high-frequency components in the signal;

[0251] Calculate the stress wave energy SE:

[0252]

[0253] Calculate the pulse energy PE:

[0254] PE=(max(|y(t)·H(t)|)) 2

[0255] Calculate the maximum peak height MP:

[0256] MP = max(|y(t)·H(t)|).

[0257] See also Figure 2 According to a specific implementation of an embodiment of the present invention, the method of using a K nearest neighbor classification algorithm to distinguish a bearing inner ring fault from an outer ring fault according to an impact characteristic value and a stress wave characteristic parameter includes:

[0258] S201, collecting a training data set with known labels, ensuring that the data set contains samples of inner ring faults and outer ring faults;

[0259] S202, extracting impact characteristic values ​​and stress wave characteristic parameters for each sample, and processing the characteristics using Zscore standardization;

[0260] S203, selecting a K value through cross-validation, training a KNN classifier using a training data set, and verifying the performance of the classifier using a test data set;

[0261] S204, extracting features and standardizing the new samples, and classifying them using the trained KNN classifier.

[0262] Corresponding to the above method embodiment, see Figure 3 The embodiment of the present invention further discloses a bearing fault diagnosis device 30 based on MVMD noise reduction and fusion of shock pulses and stress waves, comprising:

[0263] The collection module 301 uses a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, and the vibration signal data includes fault information;

[0264] The decomposition module 302 decomposes the collected fault vibration signal data using a multivariate variational mode decomposition (MVMD) method to obtain a plurality of basic modal components (IMFs), each of which represents a signal feature within a different frequency range;

[0265] The comparison module 303 calculates the correlation coefficient between each IMF and the original fault vibration signal, and screens out the IMF with a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal;

[0266] The reconstruction module 304 selects the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold value, and adds these IMFs to construct a reconstructed signal to highlight the fault characteristics;

[0267] The analysis module 305 uses a shock pulse algorithm to analyze the reconstructed signal and calculates a shock characteristic value, which is used to describe the shock intensity and frequency of the fault location;

[0268] The calculation module 306 further processes the reconstructed signal using a stress wave algorithm to calculate stress wave characteristic parameters, wherein the stress wave characteristic parameters include stress wave energy, pulse energy, and maximum peak height, and describes the severity and type of the fault based on the stress wave characteristic parameters;

[0269] The distinguishing module 307 uses a K nearest neighbor classification algorithm to distinguish the inner ring fault from the outer ring fault of the bearing according to the impact characteristic value and the stress wave characteristic parameter, thereby realizing automatic identification of the fault type;

[0270] The drawing module 308 draws the envelope spectrum of the fault signal and identifies the specific location where the fault occurs by analyzing the specific frequency bands and peaks in the envelope spectrum.

[0271] See also Figure 4 The embodiment of the present invention further provides an electronic device 60, the electronic device comprising:

[0272] at least one processor; and,

[0273] a memory communicatively connected to the at least one processor; wherein,

[0274] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the bearing fault diagnosis method based on MVMD noise reduction, fusion of impact pulses and stress waves in the aforementioned method embodiment.

[0275] An embodiment of the present invention further provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable the computer to execute the aforementioned method embodiment.

[0276] An embodiment of the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the bearing fault diagnosis method based on MVMD noise reduction, fusion of impact pulses and stress waves in the aforementioned method embodiment.

[0277] Reference below Figure 4 , which shows a schematic diagram of the structure of an electronic device 60 suitable for implementing the embodiment of the present disclosure. The electronic device in the embodiment of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0278] like Figure 4 As shown, the electronic device 60 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 60 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0279] Typically, the following devices may be connected to the I / O interface 605: input devices 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 608 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 609. The communication devices 609 may allow the electronic device 60 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4 The electronic device 60 is shown with various devices, but it should be understood that it is not required to implement or have all the devices shown. More or fewer devices may be implemented or have instead.

[0280] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 609, or installed from a storage device 608, or installed from a ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0281] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A bearing fault diagnosis method based on MVMD noise reduction, fusion of shock pulse and stress wave, characterized in that: include: S1: using a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, wherein the vibration signal data includes fault information; S2: Decompose the collected fault vibration signal data using multivariate variational mode decomposition (MVMD) to obtain multiple basic modal components (IMFs), each of which represents the signal characteristics in a different frequency range; S3: Calculate the correlation coefficient between each IMF and the original fault vibration signal, and screen out the IMF with a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal; S4: Select the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold, and add these IMFs to construct a reconstructed signal to highlight the fault characteristics; S5: applying the shock pulse algorithm to analyze the reconstructed signal and calculating the shock characteristic value, which is used to describe the shock intensity and frequency of the fault location; S6: further process the reconstructed signal using a stress wave algorithm to calculate stress wave characteristic parameters, wherein the stress wave characteristic parameters include stress wave energy, pulse energy, and maximum peak height, and describe the severity and type of the fault based on the stress wave characteristic parameters; S7: Using the K nearest neighbor classification algorithm, the inner ring fault of the bearing is distinguished from the outer ring fault according to the impact characteristic value and stress wave characteristic parameters, and the fault type is automatically identified; S8: Draw the envelope spectrum of the fault signal, and confirm the specific location of the fault by analyzing the specific frequency bands and peaks in the envelope spectrum.

2. The method according to claim 1, characterized in that The vibration sensor is used to collect vibration signal data when the inner ring and outer ring of the bearing are faulty, and the vibration signal data contains fault information, including: Use vibration sensors mounted on the mechanical structure near the bearings; After the vibration signal is collected by the data acquisition system, the collected vibration signal is amplified, filtered and digitized to obtain the final collected signal.

3. The method according to claim 2, characterized in that The multivariate variational mode decomposition (MVMD) method is used to decompose the collected fault vibration signal data to obtain multiple basic modal components (IMFs), including: Import the collected multi-channel vibration signal data into the data processing software and store the data in the form of time series, with each time point corresponding to one or more human vibration signal values; Determine the MVMD decomposition layer number K and the regularization parameter α, determine the number of IMFs after decomposition based on the decomposition layer number K, and control the smoothness and sparsity of each modal component in the decomposition process based on the regularization parameter α; The objective function of MVMD is constructed to minimize the difference between the original signal and the sum of the decomposed modal components. At the same time, a regularization term is introduced to control the smoothness and sparsity of the modal components. The objective function is expressed as: Among them, x k is the original vibration signal of the kth channel, u i is the i-th modal component, α is the regularization parameter; Use numerical optimization method to solve the above objective function and get each modal component u i .

4. The method according to claim 3, characterized in that The calculation of the correlation coefficient between each IMF and the original fault vibration signal, and screening out the IMF having a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal, includes: K IMFs are obtained through multivariate variational mode decomposition, denoted as IMF1, IMF2, …, IMF K , prepare the original fault vibration signal data, denoted as x(t); For each IMF i , calculate IMF i The correlation coefficient with the original signal x(t): Where N is the length of the signal, is the mean of the original signal x(t), It is the IMF i The mean of (t); Set the correlation coefficient threshold ρ, for each IMF i ,if It is considered that the IMF is highly correlated with the original signal and the IMF is retained. i .

5. The method according to claim 4, characterized in that The selecting of the basic modal components IMFs whose correlation coefficients are greater than a preset threshold, and adding these IMFs to construct the reconstructed signal, comprises: Add the screened highly correlated IMFs to construct the reconstructed signal y(t): Where I is the index set of the screened highly correlated IMFs.

6. The method according to claim 5, characterized in that The application of the shock pulse algorithm to analyze the reconstructed signal and calculate the shock characteristic value includes: Calculate the peak factor CF included in the shock characteristic value: Calculate the peak energy PE of the signal: PE=(max(|y(t)|)) 2 Calculate the impulse value II of the signal:

7. The method according to claim 6, characterized in that The stress wave algorithm is used to further process the reconstructed signal to calculate the stress wave energy, pulse energy and maximum peak height, including: Use a high-pass filter H(t) to extract the high-frequency components in the signal; Calculate the stress wave energy SE: Calculate the pulse energy PE: PE=(max(|y(t)·H(t)|)) 2 Calculate the maximum peak height MP: MP = max(|y(t)·H(t)|).

8. The method according to claim 7, characterized in that The method of using the K nearest neighbor classification algorithm to distinguish the inner ring fault from the outer ring fault of the bearing according to the impact characteristic value and the stress wave characteristic parameter includes: Collect a training data set with known labels to ensure that the data set contains samples of inner circle faults and outer circle faults; The impact characteristic value and stress wave characteristic parameters are extracted for each sample, and the features are processed using Zscore standardization; Select the K value through cross-validation, use the training data set to train the KNN classifier, and use the test data set to verify the performance of the classifier; Perform feature extraction and standardization on the new samples, and use the trained KNN classifier for classification.

9. A bearing fault diagnosis device based on MVMD noise reduction, fusion of shock pulse and stress wave, characterized in that: include: A collection module, using a vibration sensor to collect vibration signal data when the inner ring and outer ring of the bearing fail, wherein the vibration signal data includes failure information; The decomposition module uses the multivariate variational mode decomposition (MVMD) method to decompose the collected fault vibration signal data to obtain multiple basic modal components (IMFs), each of which represents the signal characteristics within a different frequency range. The comparison module calculates the correlation coefficient between each IMF and the original fault vibration signal, and screens out the IMF with a high correlation with the fault vibration signal by comparing the correlation between each IMF and the original signal; The reconstruction module selects the fundamental modal components IMFs whose correlation coefficients are greater than a preset threshold and adds these IMFs to construct a reconstructed signal to highlight the fault characteristics; The analysis module applies the shock pulse algorithm to analyze the reconstructed signal and calculates the shock characteristic value, which is used to describe the shock intensity and frequency of the fault location; A calculation module, which further processes the reconstructed signal using a stress wave algorithm to calculate stress wave characteristic parameters, wherein the stress wave characteristic parameters include stress wave energy, pulse energy, and maximum peak height, and describes the severity and type of the fault based on the stress wave characteristic parameters; The differentiation module uses the K nearest neighbor classification algorithm to distinguish the inner ring fault from the outer ring fault of the bearing according to the impact characteristic value and stress wave characteristic parameters, thereby realizing automatic identification of the fault type; The drawing module draws the envelope spectrum of the fault signal and confirms the specific location of the fault by analyzing the specific frequency bands and peaks in the envelope spectrum.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the bearing fault diagnosis method based on MVMD noise reduction, fusion of impact pulses and stress waves as described in any one of the preceding claims 18.

Images