A bearing fault diagnosis method and system

Abstract

The present invention belongs to the technical field of bearing fault diagnosis, and specifically provides a bearing fault diagnosis method and system, wherein the method comprises: collecting the original current signal S of the motor stator winding; performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original motor stator current signal S to obtain k component signals IMF; calculating multi-scale sample entropy on the k component signals IMF to obtain the fault feature matrix of the bearing; optimizing the support vector machine parameters using the seagull optimization algorithm, and inputting the obtained characteristic matrix of the bearing fault into the support vector machine training model; inputting the test set into the support vector machine learning model, thereby outputting the bearing fault diagnosis result. The scheme introduces multi-scale sample entropy to obtain the signal characteristics after decomposition and uses the seagull optimization SVM algorithm to classify the feature data set, thereby realizing the accurate diagnosis of the motor bearing fault, having high diagnostic accuracy, and providing a reliable basis for the safe and stable operation of the equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of bearing fault diagnosis, and more specifically, to a bearing fault diagnosis method and system. Background Art

[0002] Bearings are an important component in mechanical equipment. Its main function is to support the mechanical rotating body, reduce the friction coefficient during its movement, and ensure its rotation accuracy. With the modernization of industry, mechanical equipment is increasingly developing towards high speed and high power, and the requirements for various performance of bearings are getting higher and higher. Bearings are prone to fatigue and wear, which leads to failures. Relevant research shows that among all types of motor failures, bearing failures account for about 40%, which is one of the main forms of motor failure. Therefore, it is very important to accurately and effectively diagnose motor bearings. For rolling bearing fault diagnosis, the following problems still exist:

[0003] 1. The acquisition of bearing fault signals mostly selects the vibration signal of the bearing. The characteristic information of bearing faults can basically be detected through the vibration signal, with high accuracy and precision. However, the acquisition of vibration signals requires the installation of additional sensors in the motor, which increases the system cost on the one hand, and on the other hand, sensors cannot be installed in some occasions due to space limitations.

[0004] 2. The collected signal contains a lot of noise and interference signals. How to accurately extract fault feature information becomes the key to determine the bearing fault. The variational mode decomposition (VMD) method is a signal adaptive decomposition method. As an improved empirical mode decomposition method, VMD has a solid mathematical theoretical foundation, and the noise robustness and signal separation performance have also been greatly improved. However, the modal decomposition number k and the decomposition penalty factor α in the VMD algorithm are required to be set in advance before the signal is decomposed, but due to the actual situation, it is usually difficult to set accurately. If the setting is too large or too small, it will seriously affect the detection accuracy of the decomposed signal.

[0005] 3. How to use the decomposed signal to accurately express the fault characteristics is also the key to the accuracy of bearing diagnosis. The most commonly used method is to select based on the kurtosis maximization criterion. However, kurtosis is more sensitive to transient shocks than periodic shocks. When the signal contains a large transient shock, its kurtosis value is large, but it may not contain fault characteristics, which makes it difficult to extract the signal features.

[0006] 4. For the fault diagnosis of rotating mechanical equipment, most of the traditional fault diagnosis methods are used, which are difficult to adapt to the diagnosis of complex mechanical equipment. Machine learning methods represented by support vector machines (SVM) are an effective method for fault diagnosis of complex mechanical equipment. However, the penalty factor c and kernel function parameter g in the SVM algorithm must be set in advance according to different signals, but due to the actual situation, it is usually difficult to set them accurately. Summary of the invention

[0007] The present invention aims to solve the technical problem of low fault diagnosis accuracy of rotating mechanical equipment in the prior art.

[0008] The present invention provides a bearing fault diagnosis method, comprising the following steps:

[0009] S1, collects the original current signal S of the motor stator winding;

[0010] S2, performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF;

[0011] S3, calculate the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing;

[0012] S4, using the Seagull optimization algorithm to optimize the support vector machine parameters, and input the obtained feature matrix of the bearing fault into the support vector machine training model;

[0013] S5, input the test set into the support vector machine learning model to output the bearing fault diagnosis result.

[0014] Preferably, the S2 specifically includes the following steps:

[0015] S21: Generate the initial population. Each individual contains the decomposition number k and the penalty parameter α. The (k, α) combination of different individuals is different.

[0016] S22, calculate the fitness corresponding to each group of chromosomes in the population, and reserve a location in the individual to store the best component sequence number and fitness value;

[0017] S23, select, cross over and mutate the chromosomes in the population to generate new species;

[0018] S24, determine whether the termination condition is met, if not, jump to step S23; if satisfied, output the optimal parameters (k, α).

[0019] Preferably, the VMD algorithm constructs the fault signal into a variational constraint problem, and realizes the adaptive decomposition of the signal according to its own frequency domain characteristics by solving the variational problem, so that the sum of the estimated bandwidth of each modal function is minimized. The corresponding constrained variational model is expressed as:

[0020]

[0021]

[0022] In the formula, {u k}={u1,…,u k} represents the k sub-modes obtained by decomposition; {ω k}={ω1,…,ω k} represents the center frequency of each component.

[0023] Preferably, in order to make the signal decomposition more accurate, the following augmented Lagrange function is introduced into the VMD algorithm:

[0024]

[0025] In the formula, α is the penalty parameter; λ is the Lagrange multiplier. The multiplication operator alternating direction method is used to solve the variational problem. and n+1 Find the optimal solution of the above function, where The value problem is expressed as:

[0026]

[0027] In the formula, ω k Equivalent to Equivalent to

[0028] Preferably, the specific process of the VDM algorithm can be described as follows:

[0029] Step 1: Initialization and n;

[0030] Step 2: Update u in the frequency domain according to the formula k ,ω k ;

[0031] Step 3: Update

[0032]

[0033] Step 4: When Stop iteration when

[0034] Step 5: Repeat steps 2 and 3 until the loss coefficient e is less than the set threshold ε, and the maximum number of modal decompositions K is obtained. max , that is, the K value at this time is the maximum number of modal decompositions that the motor stator signal can be decomposed into.

[0035] Preferably, S2 specifically includes: when searching for the optimal decomposition effect parameter setting of the VMD algorithm through a genetic algorithm, specifically, assuming that individual X in a certain generation of the genetic algorithm z The corresponding parameter combination is (k z , α z ) Introduce envelope entropy to express the IMF after VMD decomposition i (j) The randomness and complexity of the component. The envelope entropy of the IMF component can be expressed as:

[0036]

[0037]

[0038] In the formula, i To decompose the IMF number; p i,j is a i (j) normalized form; a i (j) IMF i (j) Envelope signal obtained after Hilbert transform adjustment.

[0039] Preferably, the S4 specifically includes:

[0040] Step 1: Divide the input data sample matrix into a training sample matrix and a test sample matrix;

[0041] Step 2: Initialize the seagull population and set the control parameters of the SOA algorithm;

[0042] Step 3: Calculate the fitness fit, the individual optimal fitness bestfit, the group optimal fitness fitbest and the average optimal fitness avgfit;

[0043] Step 4: Update the position of the seagull and calculate the current adaptability of the seagull;

[0044] Step 5: Update the distance between the current seagull and the optimal seagull;

[0045] Step 6: Update the optimal seagull position;

[0046] Step 7: Determine whether the termination condition is met. If so, output bestc, bestg to obtain the initial SVM model. If not, repeat step 4.

[0047] The present invention also provides a bearing fault diagnosis system, which is used to implement a bearing fault diagnosis method, specifically comprising:

[0048] A signal acquisition module is used to collect the original current signal S of the motor stator winding;

[0049] The signal decomposition module is used to perform adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF;

[0050] The fault model building module calculates the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing;

[0051] The fault type diagnosis module uses the Seagull optimization algorithm to optimize the support vector machine parameters, and inputs the obtained bearing fault feature matrix into the support vector machine training model; the test set is input into the support vector machine learning model to output the bearing fault diagnosis results.

[0052] The present invention also provides an electronic device, comprising a memory and a processor, wherein the processor is used to implement the steps of a bearing fault diagnosis method when executing a computer management program stored in the memory.

[0053] The present invention also provides a computer-readable storage medium on which a computer management program is stored. When the computer management program is executed by a processor, the steps of the bearing fault diagnosis method are implemented.

[0054] Beneficial effects: The present invention provides a bearing fault diagnosis method and system, wherein the method includes: collecting the original current signal S of the motor stator winding; performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original motor stator current signal S to obtain k component signals IMF; calculating multi-scale sample entropy on the k component signals IMF to obtain the fault feature matrix of the bearing; optimizing the support vector machine parameters using the seagull optimization algorithm, and inputting the obtained characteristic matrix of the bearing fault into the support vector machine training model; inputting the test set into the support vector machine learning model, thereby outputting the bearing fault diagnosis result. This solution solves the problem that the number of VMD decompositions and the penalty factor are difficult to adapt, and at the same time introduces multi-scale sample entropy to obtain the signal characteristics after decomposition and uses the seagull optimization SVM algorithm to classify the feature data set, thereby realizing accurate diagnosis of motor bearing faults, with high diagnostic accuracy, and providing a reliable basis for safe and stable operation of equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A flow chart of a bearing fault diagnosis method provided by the present invention;

[0056] Figure 2A flow chart of an adaptive variational mode decomposition of a bearing fault diagnosis method provided by the present invention;

[0057] Figure 3 A flow chart of a bearing fault diagnosis method provided by the present invention based on a support vector machine method optimized by Seagull;

[0058] Figure 4 A schematic diagram of a current probe collecting an original signal in a bearing fault diagnosis method provided by the present invention;

[0059] Figure 5 A comparison diagram of the bearing fault model recognition based on multi-scale sample entropy between the support vector machine method provided by the present invention without the Seagull optimization algorithm and the support vector machine method with the Seagull optimization algorithm;

[0060] Figure 6 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention;

[0061] Figure 7 A schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention. DETAILED DESCRIPTION

[0062] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0063] like Figures 1 to 5 As shown, the present invention provides a bearing fault diagnosis method, wherein the method includes: collecting the original current signal S of the motor stator winding; performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original motor stator current signal S to obtain k component signals IMF; calculating multi-scale sample entropy on the k component signals IMF to obtain the fault feature matrix of the bearing; optimizing the support vector machine parameters using the seagull optimization algorithm, and inputting the obtained characteristic matrix of the bearing fault into the support vector machine training model; inputting the test set into the support vector machine learning model, thereby outputting the bearing fault diagnosis result. This solution solves the problem that the number of VMD decompositions and the penalty factor are difficult to adapt, and at the same time introduces multi-scale sample entropy to obtain the signal characteristics after decomposition and uses the seagull optimization SVM algorithm to classify the feature data set, thereby realizing the accurate diagnosis of motor bearing faults, having high diagnostic accuracy, and providing a reliable basis for the safe and stable operation of the equipment.

[0064] The purpose of the invention is to provide a bearing fault diagnosis method based on adaptive variational mode decomposition of genetic algorithm and Seagull optimized support vector machine. This solution solves the problem that the number of VMD decompositions and penalty factors are difficult to adapt to each other. At the same time, multi-scale sample entropy is introduced to obtain the signal characteristics after decomposition and the Seagull optimized SVM algorithm is used to classify the feature data set, thereby realizing accurate diagnosis of motor bearing faults, having high diagnostic accuracy, and providing a reliable basis for safe and stable operation of equipment.

[0065] The object of the present invention is achieved through the following technical solutions:

[0066] A bearing fault diagnosis method, specifically a bearing fault diagnosis method based on parameter adaptive VMD and Seagull optimized SVM, comprising the following steps:

[0067] Step S1: collecting the original current signal S of the motor stator winding;

[0068] Step S2: performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF;

[0069] Step S3: Calculate the multi-scale sample entropy for the k component signal IMFs to obtain the fault feature matrix of the bearing;

[0070] Step S4: Optimize the support vector machine parameters using the Seagull optimization algorithm, and input the obtained feature matrix of the bearing fault into the support vector machine training model;

[0071] Step S5: Input the test set into the support vector machine learning model to output the bearing fault diagnosis result.

[0072] In a further solution, in step S1, a current probe sensor is used to collect a current signal S of the motor stator winding.

[0073] In a further solution, the process of adaptive variational mode decomposition (VMD) in step S2 is:

[0074] Step S21: Generate an initial population, where each individual contains the decomposition number k and the penalty parameter α, and different individuals have different combinations of (k, α);

[0075] Step S22: Calculate the fitness corresponding to each group of chromosomes in the population, and reserve a location in the individual to store the best component sequence number and fitness value;

[0076] Step S23: performing selection, crossover and mutation operations on the chromosomes in the population to generate new species;

[0077] Step S24: Determine whether the termination condition is met. If not, jump to step S23; if satisfied, output the optimal parameters (k, α).

[0078] The VMD algorithm constructs the fault signal into a variational constraint problem, and solves the variational problem to achieve adaptive decomposition of the signal according to its own frequency domain characteristics, so that the sum of the estimated bandwidth of each modal function is minimized. The corresponding constrained variational model is expressed as:

[0079]

[0080]

[0081] In the formula, {u k}={u1,…,u k} represents the k sub-modes obtained by decomposition; {ω k}={ω1,…,ω k} represents the center frequency of each component.

[0082] In order to make the signal decomposition more accurate, the following augmented Lagrange function is introduced into the algorithm:

[0083]

[0084] In the formula, α is the penalty parameter; λ is the Lagrange multiplier. The multiplication operator alternating direction method is used to solve the variational problem. and n+1 Find the optimal solution of the above function, where The value problem is expressed as:

[0085]

[0086] In the formula, ω k Equivalent to Equivalent to

[0087] Using Parseval Fourier isometry transform, the above formula can be transformed into:

[0088]

[0089] Similarly, the updated result of the center frequency is

[0090]

[0091] In the formula, Equivalent to the current remaining amount Wiener filtering result; is the center of gravity of the power spectrum of the mode function.

[0092] The VMD algorithm is a method of continuously updating in the frequency domain, and then performing an inverse Fourier transform to obtain the result in the time domain. The specific process of the VMD algorithm can be described as follows:

[0093] Step 1: Initialization and n;

[0094] Step 2: Update u in the frequency domain according to the formula k ,ω k ;

[0095] Step 3: Update

[0096]

[0097] Step 4: When Stop iteration when

[0098] Step 5: Repeat steps 2 and 3 until the loss coefficient e is less than the set threshold ε, and the maximum number of modal decompositions K is obtained. max , that is, the K value at this time is the maximum number of modal decompositions that the motor stator signal can be decomposed into.

[0099] In the preferred solution, in step S2, in order to obtain the best decomposition effect of the VMD algorithm and avoid the influence of human factors on the decomposition effect, a reasonable optimization goal needs to be determined when searching for the best decomposition effect parameter setting of the VMD algorithm through the genetic algorithm. z The corresponding parameter combination is (k z , α z ) Introduce envelope entropy to express the IMF after VMD decomposition i (j) The randomness and complexity of the component. The envelope entropy of the IMF component can be expressed as:

[0100]

[0101]

[0102] In the formula, i To decompose the IMF number; p i,j is a i (j) normalized form; a i (j) IMF i (j) Envelope signal obtained after Hilbert transform adjustment. In order to further improve the diagnostic efficiency and accuracy and make up for the deficiency that the traditional objective function can only reflect the periodic characteristics of the signal but not the impact characteristics, the kurtosis index which is more sensitive to early bearing failure is introduced. Component IMF i (j) Kurtosis index can be expressed as:

[0103]

[0104] In the formula, u i For IMF i The mean of (j); σ i For IMF i (j) Variance.

[0105] In summary, the genetic algorithm objective function is established:

[0106]

[0107] The optimization objective can be expressed as:

[0108] min{C}

[0109] stk min ≤k z ≤k max

[0110] α min ≤α z ≤α max

[0111] Its mutation fitness solution function is:

[0112]

[0113] In the formula, C max and C min Represent the maximum and minimum comprehensive objective function values ​​of individuals in the contemporary population; C z For individual X z The comprehensive objective function value.

[0114] The expressions of mutation probability and crossover probability are:

[0115]

[0116]

[0117] In the formula, fit avg is the average fitness of individuals in the contemporary population; fit z To perform crossover or mutation operation on individual X z fitness max is the maximum fitness of individuals in the contemporary population.

[0118] In a preferred solution, the multi-scale sample entropy in step S3 can effectively reflect the degree of irregularity of a system. If the irregularity of the system is higher, the corresponding entropy value is larger; on the contrary, if the regularity of the system is weaker, it has a smaller entropy value. The multi-scale sample entropy algorithm is used to calculate the decomposed IMF decomposition to obtain the characteristic matrix of different faults.

[0119] In a preferred solution, the optimization of SVM in step S4 is based on the standard SVM, and the penalty factor c and kernel function parameter g in the standard SVM are optimized using the Seagull Optimization Algorithm (SOA). Specifically, the following steps are included:

[0120] Step 1: Divide the input data sample matrix into a training sample matrix and a test sample matrix;

[0121] Step 2: Initialize the seagull population and set the control parameters of the SOA algorithm;

[0122] Step 3: Calculate the fitness fit, the individual optimal fitness bestfit, the group optimal fitness fitbest and the average optimal fitness avgfit;

[0123] Step 4: Update the position of the seagull and calculate the current adaptability of the seagull;

[0124] Step 5: Update the distance between the current seagull and the optimal seagull;

[0125] Step 6: Update the optimal seagull position;

[0126] Step 7: Determine whether the termination condition is met. If so, output bestc, bestg to obtain the initial SVM model. If not, repeat step (4).

[0127] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

[0128] This embodiment uses the inner ring, outer ring, and rolling element faults of the bearing as shown as an example to illustrate a bearing fault diagnosis method based on parameter adaptive VMD and optimized SVM.

[0129] Experimental system composition:

[0130] Construct a motor bearing fault detection experimental system, and detect the motor bearing status by replacing the motor bearing. Use a current probe to collect the motor stator winding signal, and the collected signal is input into the computer for analysis through an acquisition card.

[0131] Signal acquisition: Use current probe to collect motor stator winding signals, such as Figure 4As shown, (a): normal signal; (b): bearing rolling element fault signal; (c): bearing inner ring fault signal; (d): bearing outer ring fault signal. The collected signals are then input into the computer through the acquisition card for analysis.

[0132] Signal processing: The number of variational modal parameter decompositions k and the penalty factor α are optimized based on a genetic algorithm, and the optimized parameters are used to decompose the bearing signals collected under four non-average states.

[0133] Bearing signals under different states decomposed by adaptive variational mode decomposition method; (a): Bearing signal decomposition under normal state; (b): Bearing signal decomposition with rolling element fault; (c): Bearing signal decomposition with inner ring fault; (d): Bearing signal decomposition with outer ring fault.

[0134] According to the decomposed IMF components, the multi-scale sample entropy of the IMF components is calculated respectively, and the multi-scale sample entropy fault model matrix is ​​constructed. Multiple sets of data are collected repeatedly so that each bearing state has 200 sets of multi-scale sample entropy models. 670 sets of multi-scale sample entropy are used for SVM training, and 120 sets are used to verify the recognition accuracy of the trained model.

[0135] Optimization and recognition of support vector machines

[0136] 670 sets of multi-scale sample entropy were input into the SOA-SVM model, and the SVM parameters were optimized by SOV. The optimized bestc and bestg were used as new SVM parameters. 120 sets of multi-scale sample entropy were input into the SVM optimized by SOA, and the accuracy of the optimized model was calculated.

[0137] like Figure 5 As shown in the figure, for the above-trained SOA-optimized SVM model, the input is the test set, and the output is the corresponding fault diagnosis prediction label, which is compared with the true label. The diagnostic accuracy of the test set is calculated to be 100%. The experimental results show that this method can accurately determine the fault type corresponding to the bearing.

[0138] It can be seen from the above embodiments that the bearing fault diagnosis method based on parameter adaptive VMD and optimized SVM of the present invention can adaptively select the best VMD decomposition parameters according to the characteristics of the signal to be analyzed, which solves the problem that the VMD decomposition parameters are difficult to adapt, thereby providing more accurate signal analysis results. Further, for the multiple IMF signals after decomposition, the best IMF containing the fault characteristic frequency can be effectively found as the input of machine learning. Machine learning adopts the SOA_SVM classification algorithm, which can autonomously optimize the best parameters, thereby realizing accurate diagnosis of bearing faults and effectively improving the recognition rate of bearing faults.

[0139] The present invention also provides a bearing fault diagnosis system, which is used to implement a bearing fault diagnosis method, specifically comprising:

[0140] A signal acquisition module is used to collect the original current signal S of the motor stator winding;

[0141] The signal decomposition module is used to perform adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF;

[0142] The fault model building module calculates the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing;

[0143] The fault type diagnosis module uses the Seagull optimization algorithm to optimize the support vector machine parameters, and inputs the obtained bearing fault feature matrix into the support vector machine training model; the test set is input into the support vector machine learning model to output the bearing fault diagnosis results.

[0144] See also Figure 6 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 6 As shown, an embodiment of the present invention provides an electronic device, including a memory 1310, a processor 1320, and a computer program 1311 stored in the memory 1310 and executable on the processor 1320. When the processor 1320 executes the computer program 1311, the following steps are implemented:

[0145] S1, collects the original current signal S of the motor stator winding;

[0146] S2, performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF;

[0147] S3, calculate the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing;

[0148] S4, using the Seagull optimization algorithm to optimize the support vector machine parameters, and input the obtained feature matrix of the bearing fault into the support vector machine training model;

[0149] S5, input the test set into the support vector machine learning model to output the bearing fault diagnosis result.

[0150] See also Figure 7 Schematic diagram of an embodiment of a computer-readable storage medium provided by the present invention. Figure 7 As shown, this embodiment provides a computer-readable storage medium 1400, on which a computer program 1411 is stored. When the computer program 1411 is executed by a processor, the following steps are implemented:

[0151] S1, collects the original current signal S of the motor stator winding;

[0152] S2, performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF;

[0153] S3, calculate the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing;

[0154] S4, using the Seagull optimization algorithm to optimize the support vector machine parameters, and input the obtained feature matrix of the bearing fault into the support vector machine training model;

[0155] S5, input the test set into the support vector machine learning model to output the bearing fault diagnosis result.

[0156] The beneficial effects of the present invention are:

[0157] 1. Regarding the selection of signals, the present invention adopts the stator winding signal of the motor, which solves the problem that the vibration signal has large interference and is difficult to collect.

[0158] 2. When the VMD method decomposes the actual signal, if k is set too small, a certain mode will contain multiple components, resulting in the inability to obtain complete information; if k is set too large, the modal center frequencies will overlap, resulting in the appearance of false modal components. The present invention can adaptively select the optimal VMD decomposition parameters according to the characteristics of the signal to be decomposed, solving the problem that the VMD decomposition parameters are difficult to adapt, thereby providing more accurate signal analysis results and achieving accurate diagnosis of bearing faults.

[0159] 3. For the multiple signals after VMD decomposition, multi-scale sample entropy is calculated respectively to construct a fault model, thus laying the foundation for accurate fault diagnosis by machine learning.

[0160] 4. The standard SVM model has low bearing fault diagnosis accuracy. The present invention adopts the SOA-optimized SVM classification algorithm. The SOA_SVM model can autonomously optimize the optimal parameters to achieve accurate bearing fault diagnosis.

[0161] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0162] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0163] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0164] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0165] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0166] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0167] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A bearing fault diagnosis method, characterized in that: The following steps are involved: S1, collects the original current signal S of the motor stator winding; S2, performing adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF; S3, calculate the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing; S4, using the Seagull optimization algorithm to optimize the support vector machine parameters, and input the obtained feature matrix of the bearing fault into the support vector machine training model; S5, inputting the test set into the support vector machine learning model to output the bearing fault diagnosis result; The S4 specifically includes: Step 1: Divide the input data sample matrix into a training sample matrix and a test sample matrix; Step 2: Initialize the seagull population and set the control parameters of the SOA algorithm; Step 3: Calculate the fitness fit, the individual optimal fitness bestfit, the group optimal fitness fitbest and the average optimal fitness avgfit; Step 4: Update the position of the seagull and calculate the current adaptability of the seagull; Step 5: Update the distance between the current seagull and the optimal seagull; Step 6: Update the optimal seagull position; Step 7: Determine whether the termination condition is met. If so, output bestc, bestg to obtain the initial SVM model. If not, repeat step 4.

2. The bearing fault diagnosis method according to claim 1, characterized in that: The S2 specifically includes the following steps: S21: Generate the initial population. Each individual contains the decomposition number k and the penalty parameter α. The (k, α) combination of different individuals is different. S22, calculate the fitness corresponding to each group of chromosomes in the population, and reserve a location in the individual to store the best component sequence number and fitness value; S23, select, cross over and mutate the chromosomes in the population to generate new species; S24, determine whether the termination condition is met, if not, jump to step S23; if satisfied, output the optimal parameters (k, α).

3. The bearing fault diagnosis method according to claim 2, characterized in that: The VMD algorithm constructs the fault signal into a variational constraint problem and solves the variational problem to achieve adaptive decomposition of the signal according to its own frequency domain characteristics, so that the sum of the estimated bandwidth of each modal function is minimized. The corresponding constrained variational model is expressed as: In the formula, {u k }={u1,…,u k } represents the k sub-modes obtained by decomposition; {ω k }={ω1,…,ω k } represents the center frequency of each component.

4. The bearing fault diagnosis method according to claim 3, characterized in that: In order to make the signal decomposition more accurate, the following augmented Lagrange function is introduced in the VMD algorithm: In the formula, α is the penalty parameter; λ is the Lagrange multiplier. The multiplication operator alternating direction method is used to solve the variational problem. and n+1 Find the optimal solution of the above function, where The value problem is expressed as: In the formula, ω k Equivalent to Equivalent to 5. The bearing fault diagnosis method according to claim 1, characterized in that: The specific process of the VDM algorithm can be described as follows: Step 1: Initialization and n; Step 2: Update u in the frequency domain according to the formula k ,ω k ; Step 3: Update Step 4: When Stop iteration when Step 5: Repeat steps 2 and 3 until the loss coefficient e is less than the set threshold ε, and the maximum number of modal decompositions K is obtained. max , that is, the K value at this time is the maximum number of modal decompositions that the motor stator signal will be decomposed into.

6. The bearing fault diagnosis method according to claim 1, characterized in that: S2 specifically includes: when searching for the optimal decomposition effect parameter setting of the VMD algorithm through the genetic algorithm, specifically, assuming that individual X in a certain generation of the genetic algorithm z The corresponding parameter combination is (k z , α z ) Introduce envelope entropy to express the IMF after VMD decomposition i (j) The randomness and complexity of the component. The envelope entropy of the IMF component can be expressed as: Where i is the decomposed IMF number; p i,j is a i (j) normalized form; a i (j) IMF i (j) Envelope signal obtained after Hilbert transform adjustment.

7. A bearing fault diagnosis system, characterized in that: The system is used to implement the bearing fault diagnosis method according to any one of claims 1 to 6, and specifically includes: A signal acquisition module is used to collect the original current signal S of the motor stator winding; The signal decomposition module is used to perform adaptive variational mode decomposition VMD based on genetic algorithm on the collected original stator current signal S of the motor to obtain k component signals IMF; The fault model building module calculates the multi-scale sample entropy of the k component signal IMF to obtain the fault feature matrix of the bearing; The fault type diagnosis module uses the Seagull optimization algorithm to optimize the support vector machine parameters, and inputs the obtained bearing fault feature matrix into the support vector machine training model; the test set is input into the support vector machine learning model to output the bearing fault diagnosis results.

8. An electronic device, characterized in that: It comprises a memory and a processor, wherein the processor is used to implement the steps of the bearing fault diagnosis method according to any one of claims 1 to 6 when executing a computer management program stored in the memory.

9. A computer-readable storage medium, characterized in that: A computer management program is stored thereon, and when the computer management program is executed by the processor, the steps of the bearing fault diagnosis method according to any one of claims 1 to 6 are implemented.

Images