Motor fault diagnosis method based on VMD-AFWOA-SVM

Abstract

The invention relates to the technical field of motor fault diagnosis, in particular to a motor fault diagnosis method based on VMD-AFWOA-SVM, and the method comprises the following steps: collecting the current data of a motor, so as to obtain four types of current data of the motor; decomposing the four types of current data through VMD, and constructing a feature vector as a current data set of the motor by calculating current data energy; improving a population initialization mode, a convergence factor and a position updating formula of the WOA algorithm; a penalty factor and kernel function parameters of the SVM model are optimized based on AFWOM, and the optimized parameters are endowed to the SVM model again to construct an AFWOA-SVM classification model; and inputting the current data of the motor into the AFWOA-SVM classification model to obtain a motor fault diagnosis result corresponding to the current data of the motor. The improved whale algorithm is utilized to optimize the penalty factor and the kernel function parameter of the SVM model, and the motor fault diagnosis precision can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of motor fault diagnosis, and in particular to a motor fault diagnosis method based on VMD-AFWOA-SVM. Background Art

[0002] In the field of geological exploration, the choice of drive mode is closely related to the characteristics of exploration technology and the actual needs of equipment. Historically, hydraulic drive technology has played an important role in drilling and other address exploration operations with its high torque output and excellent control performance, especially in the face of complex and changeable geological conditions. However, with the rapid advancement of motor technology, motor drive systems have gradually emerged and become the preferred power source for many geological exploration equipment. Despite this, the special environment of geological exploration operations (for example, remote mining areas, rugged mountain roads, and arid deserts) has brought unprecedented challenges to motor operation. These extreme environments (for example, extreme temperature fluctuations, high humidity, heavy dust pollution, long-term operation fluctuations, and high-intensity operation fluctuations) together constitute a severe test of motor performance. Under such working conditions, the motor not only needs to maintain efficient and stable operation, but also needs to have the ability to resist the invasion of harsh external factors. Even though motor technology is becoming more and more mature, its operation under such extreme conditions still faces many challenges. Common performance degradation or failure phenomena such as rotor bar breakage, air gap eccentricity, and inter-turn short circuits occur. These faults not only affect the efficiency and progress of exploration operations, but may also cause irreversible damage to exploration equipment, thereby increasing operating costs and safety risks. Therefore, how to ensure the stable and reliable operation of the motor drive system in harsh environments has become an important issue that needs to be solved in the field of geological exploration. To this end, we need to provide a motor fault diagnosis method based on VMD-AFWOA-SVM.

[0003] The traditional method of motor fault diagnosis is manual (i.e. manually judging whether the motor is faulty). This method is inefficient and has large errors. Generally speaking, motor fault diagnosis is divided into two stages:

[0004] 1. Fault signal analysis and feature extraction, including: feature extraction methods based on time domain analysis and frequency domain analysis, using EMD (empirical mode decomposition) to decompose the current signal. However, when EMD decomposes the current signal, mode aliasing may occur, making it difficult to accurately describe the true characteristics of the current signal;

[0005] 2. Fault classification: Support vector machine (SVM) has become a popular classification method in the field of pattern recognition and machine learning due to its solid theoretical foundation and superior classification effect.

[0006] The performance of the SVM model is affected by several key parameters (for example, penalty factor C, kernel function parameter g). Therefore, optimizing the parameters of the SVM model is an important step to improve model performance. However, traditional optimization algorithms have problems such as low convergence accuracy and falling into local optimal solutions. Summary of the invention

[0007] The technical problem to be solved by the present invention is: in order to solve the technical problem of low accuracy of existing motor fault diagnosis, the present invention provides a motor fault diagnosis method based on VMD-AFWOA-SVM, which can optimize the parameters of the SVM model through the improvement of the motor fault algorithm, thereby improving the accuracy of motor fault diagnosis.

[0008] The technical solution adopted by the present invention to solve the technical problem is: a motor fault diagnosis method based on VMD-AFWOA-SVM, comprising the following steps:

[0009] S1. Collecting the current data of the motor to obtain four types of current data of the motor;

[0010] S2, decomposing the four types of current data in step S1 by VMD, and calculating the energy of the current data and constructing a feature vector as a current data set of the motor;

[0011] S3, improve the population initialization method, convergence factor and position update formula of WOA algorithm;

[0012] S4. Based on AFWOM, the penalty factor and kernel function parameters of the SVM model are optimized, and the optimized parameters are reassigned to the SVM model to construct the AFWOA-SVM classification model;

[0013] S5. Input the current data of the motor in step S2 into the AFWOA=SVM classification model in step S4 to obtain the motor fault diagnosis result corresponding to the motor current data.

[0014] Therefore, the VMD algorithm is used to decompose the motor current signal, which can greatly reduce the problem of modal aliasing. The characteristic vector is obtained by calculating the energy, which can well reflect the fault characteristic information of the motor. The penalty factor and kernel function parameters of the SVM model are optimized using the improved whale algorithm, so that the fault diagnosis method shows stronger accuracy and robustness, and can improve the accuracy of motor fault diagnosis.

[0015] Furthermore, in step S2, the mathematical model of VMD decomposition current data is:

[0016]

[0017] Where: f represents the original current signal, {u k}={u1,…,u k} represents the decomposed IMF components, {ω k}={ω1,…,ω k} represents the center frequency of each IMF component, Represents an estimated center frequency of the eigenmode signal mixture.

[0018] Furthermore, step S3 includes the following steps:

[0019] S3-1, introduce LHS-Logistic method to initialize the population;

[0020] S3-2, calculate the fitness of individual whales and record the position of the best individual;

[0021] S3-3, update matrix coefficients A, C;

[0022] S3-4, determine the values ​​of p and A, and update the position in the manner of surrounding the prey, searching for the prey, and capturing the prey according to different conditions;

[0023] S3-5. Calculate the fitness of individual whales again and update the position of the optimal individual;

[0024] S3-6, determine whether the current number of iterations reaches the maximum number of iterations, if reached, output the optimal value, if not reached, return to step S3-3. Therefore, in the initialization stage of the whale algorithm population, Latin hypercube is adopted and chaotic mapping is combined. Latin hypercube is adopted to ensure that individuals are evenly distributed in each dimension, increase the diversity of the population, and at the same time, the randomness and traversal of the chaotic sequence are adopted to enhance the global search capability; the penalty factor and kernel function parameters of the SVM model are optimized using the improved whale algorithm, so that the fault diagnosis method shows stronger accuracy and robustness, and can improve the accuracy of motor fault diagnosis.

[0025] Furthermore, in step S3-1, the Logistic mapping formula is:

[0026] x n+1 = r·x n ·(1-x n );

[0027] The formula for LHS-Logistic to generate a population is:

[0028]

[0029] Where: P ij represents the Latin hypercube sample point randomly generated by the jth individual in the i-th dimension, N represents the population size or population scale, a represents the proportion of the LHS initialization strategy in the population, Li , U i Respectively represent the lower and upper limits of the i-th dimension, x n represents the chaotic sequence value generated by the nth iteration, and r represents the control parameter of the chaotic mapping.

[0030] Furthermore, in step S3-3, the updating formula of the matrix coefficient A is:

[0031] A = 2γ·r1-γ;

[0032] The update formula of the matrix coefficient C is:

[0033] C = 2·r2;

[0034] Among them: r1, r2 represent random vectors between [0, 1], γ represents the convergence factor adaptively adjusted according to population diversity;

[0035] The calculation formula of population diversity Di is:

[0036]

[0037] Where: pop represents the population size, dim represents the vector dimension, Ce j Represents the value of the population center in the jth dimension;

[0038] The calculation formula of the convergence factor γ is:

[0039]

[0040] Where: β represents the weight coefficient, t represents the current number of iterations, T max Indicates the maximum number of iterations, Di max Indicates the maximum value of population diversity. Therefore, when the whale algorithm updates the matrix coefficient, the convergence factor is improved. The improved adaptive convergence factor can adjust the parameters adaptively according to the population diversity, thereby dynamically balancing the global search and local development capabilities; the introduction of adaptive weights in the whale update position formula enables the algorithm to focus more on global search in the early stage, and pay more attention to local optimization in the later stage, thereby improving the performance of the algorithm.

[0041] Furthermore, in step S3-4, the calculation formula of the updated individual position X(t+1) is:

[0042]

[0043] The calculation formula for the distance D between the current search individual and the optimal individual in the encirclement stage is:

[0044] D=|C·X best (t)-X(t)|;

[0045] The calculation formula of the distance D″ between the current individual and the randomly selected individual in the search phase is:

[0046] D″=|C·X rand (t)-X(t)|;

[0047] The calculation formula of the direct distance D′ between the current search individual and the optimal individual in the capture phase is:

[0048] D′=|X best -X(t)|;

[0049] Where: p is a random number between [0, 1], b is a constant parameter that determines the shape of the spiral, X best (t) represents the current optimal position of the whale, X rand (t) represents the position of a randomly selected whale individual, and ω represents the introduced adaptive weight;

[0050] The calculation formula of adaptive weight ω is:

[0051]

[0052] Where: η represents the parameter that controls the degree of nonlinearity.

[0053] Furthermore, in step S1, four types of current data of the motor include:

[0054] Normal current data, current data of air gap eccentricity, current data of rotor broken bars and current data of turn-to-turn short circuit.

[0055] Furthermore, in step S4, the penalty factor and kernel function parameters in the SVM model are determined by AFWOA optimization, and the SVM model is retrained and tested.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. The present invention uses the VMD algorithm to decompose the motor current signal, which can greatly reduce the problem of modal aliasing. By calculating the energy to obtain the characteristic vector, it can well reflect the fault characteristic information of the motor.

[0058] 2. The present invention combines Latin hypercube and chaotic mapping in the population initialization stage of the whale algorithm. Latin hypercube ensures that individuals are evenly distributed in each dimension, increasing the diversity of the population. At the same time, the randomness and traversal of chaotic sequences are used to enhance the global search capability.

[0059] 3. The present invention improves the convergence factor when the whale algorithm updates the matrix coefficients. The improved adaptive convergence factor can adaptively adjust parameters according to population diversity, thereby dynamically balancing global search and local development capabilities.

[0060] 4. The present invention introduces adaptive weights into the whale update position formula, so that the algorithm can focus more on global search in the early stage and pay more attention to local optimization in the later stage, thereby improving the performance of the algorithm.

[0061] 5. The present invention utilizes the improved whale algorithm to optimize the penalty factor and kernel function parameters of the SVM model, so that the fault diagnosis method exhibits stronger accuracy and robustness, and can improve the accuracy of motor fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0063] Figure 1 It is a structural schematic diagram of the motor fault diagnosis method based on VMD-AFWOA-SVM of the present invention;

[0064] Figure 2 It is a comparison chart of the test results of the AFWOA algorithm of the present invention and the existing algorithm;

[0065] Figure 3 This is the fault classification result diagram of the EMD-SVM of the present invention;

[0066] Figure 4 This is a fault classification result diagram of the VMD-SVM of the present invention;

[0067] Figure 5 This is a fault classification result diagram of the VMD-WOA-SVM of the present invention;

[0068] Figure 6 This is a diagram of the fault classification results of the VMD-AFWOA-SVM of the present invention. DETAILED DESCRIPTION

[0069] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0070] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0071] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0072] like Figures 1 to 6 As shown, it is the best embodiment of the present invention. The motor fault diagnosis method based on VMD-AFWOA-SVM in this embodiment includes the following steps:

[0073] S1. Collecting the current data of the motor to obtain four types of current data of the motor;

[0074] S2, decomposing the four types of current data in step S1 by VMD, and calculating the energy of the current data and constructing a feature vector as a current data set of the motor;

[0075] S3, improve the population initialization method, convergence factor and position update formula of WOA algorithm;

[0076] S4. Based on AFWOM, the penalty factor and kernel function parameters of the SVM model are optimized, and the optimized parameters are reassigned to the SVM model to construct the AFWOA-SVM classification model;

[0077] S5, input the current data of the motor in step S2 into the AFWOA-SVM classification model in step S4 to obtain the motor fault diagnosis result corresponding to the motor current data. Therefore, the use of VMD algorithm to decompose the motor current signal can greatly reduce the problem of modal aliasing, and obtain the characteristic vector by calculating the energy, which can well reflect the fault characteristic information of the motor; the penalty factor and kernel function parameters of the SVM model are optimized by using the improved whale algorithm, so that the fault diagnosis method shows stronger accuracy and robustness, and can improve the accuracy of motor fault diagnosis.

[0078] Specifically, in step S1, during collection, it is necessary to set the sampling frequency and sampling time.

[0079] Specifically, in step S2, the number of decomposition modes is set to 5 and the penalty factor is 2000.

[0080] Specifically, before step S3, according to the calculation formula Calculate the energy of each IMF component and construct the eigenvector T = [Ec1, Ec2, ..., Ec i ];

[0081] Where: n is the number of sampling points of an IMF component, x jk is the amplitude of the jth IMF component at the kth point;

[0082] In addition, category labels are added to the feature vector dataset, the training set and test set are divided in a ratio of 3:2, and the data are normalized.

[0083] Specifically, in step S4, the obtained optimal parameters are input into the SVM model and the training set is input for training; the RBF kernel function is introduced for the distribution characteristics of the input data, and at the same time, a one-to-many strategy is applied to multi-fault diagnosis to produce multiple support vector machine binary classification models; new input data samples are received and classified through the trained support vector machine model, and the category of the data is predicted based on the decision-making convenient results.

[0084] In this embodiment, in step S1 , the four types of current data of the motor include: normal current data, current data of air gap eccentricity, current data of rotor bar breakage, and current data of turn-to-turn short circuit.

[0085] In this embodiment, in step S2, the mathematical model of VMD decomposition current data is:

[0086]

[0087] Where: f represents the original current signal, {u k}={u1,…,u k} represents the decomposed IMF components, {ω k}={ω1,…,ω k} represents the center frequency of each IMF component, Represents an estimated center frequency of the eigenmode signal mixture.

[0088] In this embodiment, step S2 includes the following steps:

[0089] S2-1. Initialization λ 1 and n = 0, setting the decomposition mode number k and the penalty factor α;

[0090] S2-2, number of iterations N = n + 1;

[0091] S2-3, combining the alternating method multiplier method and Fourier equidistant change, alternating update and n+1 ;

[0092] S2-4, point accuracy ε, if the stop condition is met Stop the loop;

[0093] In step S2-3, the representation of the kth component in the frequency domain at the n+1th iteration is for:

[0094]

[0095] Representation of the center frequency corresponding to the kth component at the n+1th iteration for:

[0096]

[0097] Representation of Lagrange multipliers in the n+1th cycle in the frequency domain for:

[0098]

[0099] in: is the representation of the original signal in the frequency domain.

[0100] In this embodiment, step S3 includes the following steps:

[0101] S3-1, introduce LHS-Logistic method to initialize the population;

[0102] S3-2, calculate the fitness of individual whales and record the position of the best individual;

[0103] S3-3, update matrix coefficients A, C;

[0104] S3-4, determine the values ​​of p and A, and update the position in the manner of surrounding the prey, searching for the prey, and capturing the prey according to different conditions;

[0105] S3-5. Calculate the fitness of individual whales again and update the position of the optimal individual;

[0106] S3-6, determine whether the current number of iterations reaches the maximum number of iterations, if so, output the optimal value, if not, return to step S3-3;

[0107] In step S3-1, the Logistic mapping formula is:

[0108] x n+1 = r·x n ·(1-x n );

[0109] The formula for LHS-Logistic to generate a population is:

[0110]

[0111] Where: P ij represents the Latin hypercube sample point randomly generated by the jth individual in the i-th dimension, N represents the population size or population scale, a represents the proportion of the LHS initialization strategy in the population, L i , U i Respectively represent the lower and upper limits of the i-th dimension, x n represents the chaotic sequence value generated by the nth iteration, r represents the control parameter of the chaotic mapping (usually 3.99);

[0112] In step S3-3, the update formula of the matrix coefficient A is:

[0113] A = 2γ·r1-γ;

[0114] The update formula of the matrix coefficient C is:

[0115] C = 2·r2;

[0116] Among them: r1, r2 represent random vectors between [0, 1], γ represents the convergence factor adaptively adjusted according to population diversity;

[0117] The calculation formula of population diversity Di is:

[0118]

[0119] Where: pop represents the population size, dim represents the vector dimension, Ce j Represents the value of the population center in the jth dimension;

[0120] The calculation formula of the convergence factor γ is:

[0121]

[0122] Where: β represents the weight coefficient, t represents the current number of iterations, T max Indicates the maximum number of iterations, Di max Indicates the maximum population diversity;

[0123] In step S3-4, the calculation formula of the updated individual position X(t+1) is:

[0124]

[0125] The calculation formula for the distance D between the current search individual and the optimal individual in the encirclement stage is:

[0126] D=|C·X best (t)-X(t)|;

[0127] The calculation formula of the distance D″ between the current individual and the randomly selected individual in the search phase is:

[0128] D″=|C·X rand (t)-X(t)|;

[0129] The calculation formula of the direct distance D′ between the current search individual and the optimal individual in the capture phase is:

[0130] D′=|X best -X(t)|;

[0131] Where: p is a random number between [0, 1], b is a constant parameter that determines the shape of the spiral, X best (t) represents the current optimal position of the whale, X rand (t) represents the position of a randomly selected whale individual, and ω represents the introduced adaptive weight;

[0132] The calculation formula of adaptive weight ω is:

[0133]

[0134] Where: η is a parameter that controls the degree of nonlinearity (set to 0.4).

[0135] Specifically, in step S3-1, the final population individuals are composed of individuals generated by two methods. Therefore, in the initialization stage of the whale algorithm population, Latin hypercube is adopted and chaotic mapping is combined. Latin hypercube is adopted to ensure that individuals are evenly distributed in each dimension, increasing the diversity of the population. At the same time, the randomness and traversal of the chaotic sequence are adopted to enhance the global search capability; the penalty factor and kernel function parameters of the SVM model are optimized by using the improved whale algorithm, so that the fault diagnosis method shows stronger accuracy and robustness, and can improve the accuracy of motor fault diagnosis; when the whale algorithm updates the matrix coefficient, the convergence factor is improved, and the improved adaptive convergence factor can adjust the parameters adaptively according to the population diversity, thereby dynamically balancing the global search and local development capabilities; the introduction of adaptive weights in the whale update position formula enables the algorithm to be more global in the early stage, and pay more attention to local optimization in the later stage, thereby improving the performance of the algorithm;.

[0136] Specifically, the settings of the improved WOA algorithm parameters are shown in Table 1.

[0137] Table 1: Parameters of the improved WOA algorithm

[0138] parameter Settings Population size 30 Iterations 20 Penalty Factor C [0.001，100] Kernel function parameter g [0.001，100]

[0139] Specifically, in step S3-2, the accuracy is used as the fitness of the algorithm, and the calculation formula of the accuracy is:

[0140]

[0141] In this embodiment, in step S4, the penalty factor and kernel function parameters in the SVM model are determined by AFWOA optimization, and the SVM model is retrained and tested.

[0142] It should be noted that: 1. Figure 2 As shown in the figure, in order to verify the effectiveness of the motor fault diagnosis method, the same test function is optimized with the WOA algorithm and the SSA algorithm. The population size is set to 30 and the number of iterations is 500. In order to see the performance comparison more intuitively, the fitness value is transformed by log 10. It can be seen that the optimization and convergence performance of the AFWOA algorithm are better than those of WOA and SSA.

[0143] 2. Collect the current signal of the motor for 5.5S, the sampling frequency is 25KHZ, the motor model is Y160M-4, the rated speed is 1460r / min, and the rated power is 11KW;

[0144] The VMD decomposition is performed on the current data, and then the energy is calculated to construct the eigenvectors. Some eigenvectors are shown in Table II;

[0145] Table 2: Representation of some feature vectors

[0146]

[0147]

[0148] 3. The classification labels of motor fault types are shown in Table 3;

[0149] Table 3: Classification of motor fault types

[0150] Fault type Label normal 1 Air gap eccentricity 2 Broken rotor bars 3 Turn-to-turn short circuit 4

[0151] 4. Normalize the data set, a total of 1000 groups of samples, divide the training set and the test set in a ratio of 3:2, that is, 150 groups of samples in each state are used as training samples, a total of 600 groups of samples are input into AWOA-SVM for parameter optimization, and 400 groups are used as test samples and input into the optimal model for testing;

[0152] 5. If Figures 3 to 6 As shown in the figure, the classification results of the test set are respectively obtained by using the models EMD-SVM, VMD-SVM, VMD-WOA-SVM, and VMD-AFWOA-SVM. The classification accuracy of the motor fault diagnosis method (i.e., VMD-AFWOA-SVM) is 99%, while the classification accuracy of VMD-WOA-SVM, VMD-SVM, and EMD-SVM are 98.25%, 96.25%, and 92.25%, respectively. Therefore, the model proposed in this motor fault diagnosis method has better superiority.

[0153] In summary, the present invention uses VMD algorithm to decompose the motor current signal, which can greatly reduce the problem of modal aliasing. By calculating the energy to obtain the characteristic vector, it can well reflect the fault characteristic information of the motor; in the initialization stage of the whale algorithm population, Latin hypercube and chaotic mapping are combined. Latin hypercube ensures that individuals are evenly distributed in each dimension, increasing the diversity of the population. At the same time, the randomness and traversal of the chaotic sequence are used to enhance the global search capability; when the whale algorithm updates the matrix coefficient, the convergence factor is improved, and the improved adaptive convergence factor can adaptively adjust the parameters according to the population diversity, thereby dynamically balancing the global search and local development capabilities; adaptive weights are introduced in the whale update position formula, so that the algorithm can be more global in the early stage, and pay more attention to local optimization in the later stage, thereby improving the performance of the algorithm; the penalty factor and kernel function parameters of the SVM model are optimized using the improved whale algorithm, so that the fault diagnosis method shows stronger accuracy and robustness, and can improve the accuracy of motor fault diagnosis.

[0154] The above description is based on the ideal embodiment of the present invention. Through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A motor fault diagnosis method based on VMD-AFWOA-SVM, characterized in that: The following steps are involved: S1. Collecting the current data of the motor to obtain four types of current data of the motor; S2, decomposing the four types of current data in step S1 by VMD, and calculating the energy of the current data and constructing a feature vector as a current data set of the motor; S3, improve the population initialization method, convergence factor and position update formula of WOA algorithm; S4. Based on AFWOM, the penalty factor and kernel function parameters of the SVM model are optimized, and the optimized parameters are reassigned to the SVM model to construct the AFWOA-SVM classification model; S5. Input the current data of the motor in step S2 into the AFWOA-SVM classification model in step S4 to obtain the motor fault diagnosis result corresponding to the motor current data.

2. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 1 is characterized in that: In step S2, the mathematical model of VMD decomposition current data is: Where: f represents the original current signal, {u k }={u1,…,u k } represents the decomposed IMF components, {ω k }={ω1,…,ω k } represents the center frequency of each IMF component, Represents an estimated center frequency of the eigenmode signal mixture.

3. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 1 is characterized in that: The step S3 comprises the following steps: S3-1, introduce LHS-Logistic method to initialize the population; S3-2, calculate the fitness of individual whales and record the position of the best individual; S3-3, update matrix coefficients A, C; S3-4, determine the values ​​of p and A, and update the position in the manner of surrounding the prey, searching for the prey, and capturing the prey according to different conditions; S3-5. Calculate the fitness of individual whales again and update the position of the optimal individual; S3-6. Determine whether the current number of iterations reaches the maximum number of iterations. If so, output the optimal value. If not, return to step S3-3.

4. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 3 is characterized in that: In step S3-1, the Logistic mapping formula is: x n+1 =r·x n ·(1-x n ); The formula for LHS-Logistic to generate a population is: Where: P ij represents the Latin hypercube sample point randomly generated by the jth individual in the i-th dimension, N represents the population size or population scale, a represents the proportion of the LHS initialization strategy in the population, L i , U i Respectively represent the lower and upper limits of the i-th dimension, x n represents the chaotic sequence value generated by the nth iteration, and r represents the control parameter of the chaotic mapping.

5. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 1 is characterized in that: In step S3-3, the update formula of the matrix coefficient A is: A = 2γ·r1-γ; The update formula of the matrix coefficient C is: C＝2·r2； Among them: r1, r2 represent random vectors between [0, 1], γ represents the convergence factor adaptively adjusted according to population diversity; The calculation formula of population diversity Di is: Where: pop represents the population size, dim represents the vector dimension, Ce j Represents the value of the population center in the jth dimension; The calculation formula of the convergence factor γ is: Where: β represents the weight coefficient, t represents the current number of iterations, T max Indicates the maximum number of iterations, Di max Indicates the maximum population diversity.

6. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 1 is characterized in that: In step S3-4, the calculation formula of the updated individual position X(t+1) is: The calculation formula for the distance D between the current search individual and the optimal individual in the encirclement stage is: D=|C·X best (t)-X(t)|; The calculation formula of the distance D″ between the current individual and the randomly selected individual in the search phase is: D″=|C·X rand (t)-X(t)|; The calculation formula of the direct distance D' between the current search individual and the optimal individual in the capture phase is: D'=|X best -X(t)|; Where: p is a random number between [0, 1], b is a constant parameter that determines the shape of the spiral, X best (t) represents the current optimal position of the whale, X rand (t) represents the position of a randomly selected whale individual, and ω represents the introduced adaptive weight; The calculation formula of adaptive weight ω is: Where: η represents the parameter that controls the degree of nonlinearity.

7. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 1 is characterized in that: In step S1, four types of current data of the motor include: Normal current data, current data of air gap eccentricity, current data of rotor broken bars and current data of turn-to-turn short circuit.

8. The motor fault diagnosis method based on VMD-AFWOA-SVM according to claim 1 is characterized in that: In step S4, the penalty factor and kernel function parameters in the SVM model are determined by AFWOA optimization, and the SVM model is retrained and tested.

Images