A new intelligent rotating machinery fault diagnosis method

Abstract

The application provides a novel intelligent rotating machinery fault diagnosis method, and belongs to the technical field of rotating machinery fault diagnosis, and the content comprises the following steps: obtaining vibration signals of rotating machinery under normal and different fault types; adopting a variational mode decomposition method to pretreat the rotating machinery vibration signals, decomposing and removing noise, and then reconstructing a new signal; adopting a short-time Fourier transform to convert the new vibration signal into a two-dimensional time-frequency image; adopting a Sahara silver ant optimization algorithm (SSAO) to optimize hyperparameters of a convolutional neural network (CNN)-long short-term memory neural network (BiLSTM), including the number of convolutional layer neurons, the size of a convolution kernel, and the number of BiLSTM layer neurons; inputting the two-dimensional time-frequency image into the CNN-BiLSTM model with the optimized hyperparameters to realize rotating machinery fault diagnosis; and the application has very important practical significance for improving the rotating machinery fault diagnosis accuracy and guiding equipment maintenance and repair.

Description

Technical Field

[0001] This invention relates to the field of rotating machinery fault diagnosis technology, and more specifically, to a novel intelligent method for diagnosing rotating machinery faults. Background Technology

[0002] Rotating machinery is a core piece of equipment in industrial production, and its operating status directly affects the stability of the production system. Equipment failures can range from minor production interruptions and reduced product quality to major safety accidents and huge economic losses. Therefore, building an efficient fault diagnosis system has significant engineering value.

[0003] In recent years, deep learning technology has made significant progress in the field of rotating machinery fault diagnosis due to its powerful feature learning capabilities. Mainstream methods convert vibration signals into time-frequency images using time-frequency analysis techniques, and then employ convolutional neural networks and their improved architectures for feature extraction and fault classification. However, the diagnostic performance of existing methods is highly dependent on the optimized configuration of network hyperparameters, and traditional manual parameter tuning methods have limitations. To address these issues, systematically exploring the optimal hyperparameter combination by constructing a parameter optimization objective function and combining it with intelligent optimization algorithms has become a research hotspot. However, existing intelligent optimization algorithms still have room for further improvement in terms of optimization accuracy and computational efficiency. Summary of the Invention

[0004] The main objective of this invention is to address the shortcomings of existing technologies and provide a novel intelligent method for diagnosing faults in rotating machinery. First, variational mode decomposition (VMD) is used for signal denoising preprocessing. Then, a convolutional neural network-bidirectional long short-term memory network model (CNN-BiLSTM) optimized using the Saharan Silver Ant Optimization (SSAO) algorithm is established. This invention originally proposes a novel Saharan Silver Ant Optimization (SSAO) algorithm and uses it to optimize the hyperparameters of the CNN-BiLSTM, resulting in a lower iterative loss value and higher diagnostic accuracy, providing reliable technical support for improving the accuracy of fault diagnosis in rotating machinery.

[0005] In nature, Sahara silver ants have evolved highly efficient cooperative foraging strategies in the extremely hot desert environment. When the temperature exceeds a critical threshold, the ant colony adopts a dynamic team formation mechanism to form cooperative squads, with individuals constantly sharing information to optimize foraging paths. In team behavior, experienced individuals gradually evolve into leaders, responsible for deciding whether to steadily advance along the current optimal path or continue exploring near areas with concentrated food resources. The behavior of team members is influenced by the leader: when individuals approach a food source, they usually follow the path left by their team leader and perform local exploration along the path; when the distance is far, they tend to quickly approach the leader's position and coordinate the search. In addition, when local information is limited or the surrounding environment changes drastically, some individuals may abandon the guidance of their original leader and instead follow a random leader from another squad to explore. Through continuous information exchange and path adjustment, the ant colony maintains efficient cooperation and foraging capabilities in high-temperature environments. The Sahara silver ant optimization algorithm proposed in this invention is a metaheuristic swarm intelligence optimization algorithm created based on the above natural phenomena.

[0006] This invention is implemented as follows:

[0007] The technical solution to achieve the purpose of this invention is: a novel intelligent method for diagnosing faults in rotating machinery, comprising the following steps:

[0008] Step 1: Obtain the vibration signals and fault classifications of the rotating machinery under normal conditions and different fault types;

[0009] Step 2: Variational mode decomposition is used to perform noise reduction preprocessing on the vibration signal to be trained, and the reconstructed signal after preprocessing is obtained;

[0010] Step 3: Convert the reconstructed signal into a two-dimensional time-frequency image using a short-time Fourier transform;

[0011] Step 4: Set the parameters required for the Sahara Silver Ant optimization algorithm;

[0012] Step 5: Optimize the hyperparameters of the convolutional neural network-bidirectional long short-term memory network based on the Sahara silver ant optimization algorithm;

[0013] Step 6: Based on the optimized hyperparameters, reconstructed signals, and fault classification, train a rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM;

[0014] Step 7: Denoise the fault diagnosis signal to be diagnosed, convert it into a two-dimensional time-frequency image through short-time Fourier transform, and input the image into the trained rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM to realize rotating machinery fault diagnosis.

[0015] Furthermore, in step two, the noise reduction preprocessing includes the following steps: setting the number of intrinsic mode functions (IMFs) and penalty factor required for variational mode decomposition (VMD) noise reduction; decomposing the original signal using a variational mode decomposition algorithm to obtain multiple IMF components; arranging the IMFs in descending order of their center frequencies and calculating the correlation of each IMF; marking the IMF with the lowest correlation as a noisy IMF; removing all IMFs marked as noise and summing the remaining IMFs to obtain the denoised reconstructed signal.

[0016] Furthermore, in step four, the Sahara silver ant optimization algorithm consists of five steps executed sequentially:

[0017] Step S41: In the initialization phase, generate using the Latin hypercube sampling method. M The initial distribution of silver ants within the thermal region can be expressed mathematically as follows:

[0018] X k = lb +1 / M ×( Y + μ- 1)( ub - lb )

[0019] in, X k Indicates the first k The location of a single silver ant; lb and ub These are the lower and upper bounds of the search space, respectively. Y For {1,2,…, M The random permutation of} μ This represents a random number taken from the interval (0,1);

[0020] Step S42: During the thermal adaptation and formation phase, silver ants will randomly form teams to update their positions, with the number of members in each team being... Ts It is dynamically adjusted, as follows:

[0021] Ts =max(3,min(5,round(3+2sin(2π× t / T ))))

[0022] The `round()` function rounds the value within the parentheses to the nearest integer. t For the current iteration, T This represents the total number of iterations.

[0023] Step S43: In the leader election phase, by evaluating each team's... Ts The fitness value of each silver ant is used to select the individual with the best fitness to serve as the leader of the team;

[0024] Step S44: In the ant path optimization phase, for each team leader, based on their fitness... F i,b t Fitness with the current global optimum F g t The proportional relationship between them determines the search strategy: if F i,b t < β × F g t Develop a path along the direction leading to the global optimum, where β These are the behavior control parameters; otherwise, local exploration is performed directly near the global optimal location, as follows:

[0025] β =1+1 / 2× U (0,1)

[0026] X i,b t+1 = X i,b t +| N (0,1)|×( X g t - X i,b t )+ N (0,1)× τ ( n )× P F i,b t < β × F g t

[0027] X i,b t+1 = X g t ×( U (0,1) +τ ( n ) ×P ) other

[0028] in, U(0,1) represents a uniformly distributed random variable in the interval [0,1]. F i,b t Representing the t In the nth iteration i The fitness score of each team leader F g t Representing the t The global optimal solution in the next iteration; X i,b t+1 and X i,b t They represent the first t+ 1st iteration and the 1st iteration t In the nth iteration i The position of a team leader X g t Indicates the first t The global optimal position in the next iteration. F i,b t and F g t These represent the fitness values ​​for the team leader and the globally optimal position, respectively, with superscripts indicating the fitness values. t Represents the current iteration number. N (0,1) is a standard normally distributed random variable. τ ( n (to obey the degree of freedom) n A random variable with a T-distribution, wherein n =5+round(10×(1–( t / T The round() function rounds the value within the parentheses to the nearest integer. P=(X k t -lb) / (ub-lb) ,in X k t Indicates the first k Only one silver ant individual in the first t Position in the next iteration;

[0029] Step S45: During the exploration phase in unknown areas, for each team member, based on their fitness level... F i,m t Compatibility with team leaders F i,b t The proportional relationship between them determines the exploration strategy: if Fi,m t <β×F i,b t If the search is successful, a cross-team learning approach is adopted, incorporating information from the best individuals in other teams for the search; otherwise, local exploration is performed directly near the current team leader. The update formula for this stage is as follows:

[0030] X i,m t+1 =rand ( X i,b t ) + | N (0,1)| × ( rand ( X i,b t ) -X i,m t ) F i,m t <β×F i,b t

[0031] X i,m t+1 = X i,b t ×( U (0,1) +τ ( n ) ×P ) other

[0032] in, X i,m t+1 and X i,m t They represent the first t +1 iterations and the t In the next iteration, the team i The Middle m The position of each member rand ( X i,b t ) indicates from the first t The position of a leader randomly selected from the team leaders in the next iteration;

[0033] After the five stages of calculation are completed, it is determined whether the iteration is complete. If the iteration is complete, the search ends and the optimal position is output. If the iteration is not complete, the process returns to step S42 and repeats steps S42 to S45.

[0034] Furthermore, in step four, the parameters required for the Sahara silver ant optimization algorithm include the total number of individuals. M The largest team Ts max Minimum size Ts min Maximum number of iterations T .

[0035] Furthermore, in step five, the Sahara Silver Ant optimization algorithm optimizes the hyperparameters of CNN-BiLSTM, including the number of neurons in the convolutional layer, the size of the convolutional kernel, and the number of neurons in the long short-term memory layer.

[0036] Furthermore, in step six, the CNN-BiLSTM model includes four convolutional layers, three pooling layers, one flattening layer, one long short-term memory layer, and one output layer. The activation function of the convolutional layers is the ReLU function, the stride is (2, 2), and zero padding is used. The pooling layer size is 2×2. The output layer outputs 4 cells, and the activation function is the softmax function. The number of neurons in the convolutional layers, the size of the convolutional kernels, and the number of neurons in the long short-term memory layer are obtained from step five.

[0037] The beneficial effects of this invention are as follows: This invention proposes a novel metaheuristic swarm intelligence optimization algorithm—the Sahara Silver Ant Cooperative Optimization Algorithm (SSAO). Through dynamic team formation and multi-source information sharing mechanisms, individuals continuously transmit environmental and location information, enabling search behavior to focus on the vicinity of the individual with the best fitness, thereby improving optimization efficiency. Furthermore, unlike other algorithms, this invention introduces a cross-team leader switching strategy under specific conditions, allowing individuals to abandon their original team leader and instead follow a random leader from another team for exploration, effectively enhancing population diversity and helping to avoid getting trapped in local optima. Based on the SSAO algorithm, this invention optimizes the hyperparameters of CNN-BiLSTM, enabling the optimized CNN-BiLSTM to extract fault features more effectively and improving fault diagnosis accuracy. Attached Figure Description

[0038] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0039] Figure 1This is a flowchart of a novel intelligent rotating machinery fault diagnosis method provided by an embodiment of the present invention;

[0040] Figure 2 This is a detailed flowchart of the SSAO optimization algorithm provided in the embodiments of the present invention;

[0041] Figure 3 This is a topology diagram of the CNN-BiLSTM network model;

[0042] Figure 4 This is a confusion matrix diagram of the fault diagnosis results of the SSAO-VMD-CNN-BiLSTM method;

[0043] Figure 5 This is a comparison chart of iterative losses for different methods;

[0044] Figure 6 This is a comparison chart of the accuracy of different iterative methods. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the implementation examples and accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to represent selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] The method of the present invention will be illustrated below using a specific case of rotating machinery fault diagnosis as an example.

[0047] like Figure 1 A novel intelligent method for diagnosing faults in rotating machinery includes the following steps:

[0048] Step 1: Obtain the vibration signals and fault classifications of the rotating machinery under normal conditions and different fault types;

[0049] Step 2: Variational mode decomposition is used to perform noise reduction preprocessing on the vibration signal to be trained, and the reconstructed signal after preprocessing is obtained;

[0050] Step 3: Convert the reconstructed signal into a two-dimensional time-frequency image using a short-time Fourier transform;

[0051] Step 4: Set the parameters required for the Sahara Silver Ant optimization algorithm;

[0052] Step 5: Optimize the hyperparameters of the convolutional neural network-bidirectional long short-term memory network based on the Sahara silver ant optimization algorithm;

[0053] Step 6: Based on the optimized hyperparameters, reconstructed signals, and fault classification, train a rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM;

[0054] Step 7: Denoise the fault diagnosis signal to be diagnosed, convert it into a two-dimensional time-frequency image through short-time Fourier transform, and input the image into the trained rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM to realize rotating machinery fault diagnosis.

[0055] Furthermore, the denoising preprocessing includes the following steps: setting the number of intrinsic mode functions (IMFs) and penalty factor required for variational mode decomposition (VMD) denoising; decomposing the original signal using a variational mode decomposition algorithm to obtain multiple IMF components; arranging the IMFs in descending order of their center frequencies and calculating the correlation of each IMF; marking the IMF with the lowest correlation as a noisy IMF; removing all IMFs marked as noise and summing the remaining IMFs to obtain the denoised reconstructed signal.

[0056] Furthermore, the Sahara silver ant optimization algorithm consists of five sequentially executed steps:

[0057] Step S41: In the initialization phase, generate using the Latin hypercube sampling method. M The initial distribution of silver ants within the thermal region can be expressed mathematically as follows:

[0058] X k = lb +1 / M ×( Y + μ- 1)( ub - lb )

[0059] in, X k Indicates the first k The location of a single silver ant; lb and ub These are the lower and upper bounds of the search space, respectively. Y For {1,2,…, M The random permutation of} μ This represents a random number taken from the interval (0,1);

[0060] Step S42: During the thermal adaptation and formation phase, silver ants will randomly form teams to update their positions, with the number of members in each team being... Ts It is dynamically adjusted, as follows:

[0061] Ts =max(3,min(5,round(3+2sin(2π× t / T ))))

[0062] The `round()` function rounds the value within the parentheses to the nearest integer. t For the current iteration, T This represents the total number of iterations.

[0063] Step S43: In the leader election phase, by evaluating each team's... Ts The fitness value of each silver ant is used to select the individual with the best fitness to serve as the leader of the team;

[0064] Step S44: In the ant path optimization phase, for each team leader, based on their fitness... F i,b t Fitness with the current global optimum F g t The proportional relationship between them determines the search strategy: if F i,b t < β × F g t Develop a path along the direction leading to the global optimum, where β These are the behavior control parameters; otherwise, local exploration is performed directly near the global optimal location, as follows:

[0065] β =1+1 / 2× U (0,1)

[0066] X i,b t+1 = X i,b t +| N (0,1)|×( X g t - X i,b t )+ N (0,1)× τ ( n)× P F i,b t < β × F g t

[0067] X i,b t+1 = X g t ×( U (0,1) +τ ( n ) ×P ) other

[0068] in, U (0,1) represents a uniformly distributed random variable in the interval [0,1]. F i,b t Representing the t In the nth iteration i The fitness score of each team leader F g t Representing the t The global optimal solution in the next iteration; X i,b t+1 and X i,b t They represent the first t+ 1st iteration and the 1st iteration t In the nth iteration i The position of a team leader X g t Indicates the first t The global optimal position in the next iteration. F i,b t and F g t These are the fitness values ​​for the team leader and the globally optimal position, respectively, with superscripts indicating their respective fitness values. t Represents the current iteration number. N (0,1) is a standard normally distributed random variable. τ ( n (to obey the degree of freedom) n A random variable with a T-distribution, wherein n =5+round(10×(1–( t / TThe round() function rounds the value within the parentheses to the nearest integer. P=(X k t -lb) / (ub-lb) ,in X k t Indicates the first k Only one silver ant individual in the first t Position in the next iteration;

[0069] Step S45: During the exploration phase in unknown areas, for each team member, based on their fitness level... F i,m t Compatibility with team leaders F i,b t The proportional relationship between them determines the exploration strategy: if F i,m t <β×F i,b t If the search is successful, a cross-team learning approach is adopted, incorporating information from the best individuals in other teams for the search; otherwise, local exploration is performed directly near the current team leader. The update formula for this stage is as follows:

[0070] X i,m t+1 =rand ( X i,b t ) + | N (0,1)| × ( rand ( X i,b t ) -X i,m t ) F i,m t <β×F i,b t

[0071] X i,m t+1 = X i,b t ×( U (0,1) +τ ( n ) ×P ) other

[0072] in, X i,m t+1 and X i,m t They represent the first t +1 iterations and the t In the next iteration, the team i The Middle m The position of each member rand ( X i,b t ) indicates from the first t The position of a leader randomly selected from the team leaders in the next iteration;

[0073] After the five stages of calculation are completed, it is determined whether the iteration is complete. If the iteration is complete, the search ends and the optimal position is output. If the iteration is not complete, the process returns to step S42 and repeats steps S42 to S45.

[0074] Furthermore, in step four, the parameters required for the Sahara silver ant optimization algorithm include the total number of individuals. M The largest team Ts max Minimum size Ts min Maximum number of iterations T .

[0075] Furthermore, in step five, the Sahara Silver Ant optimization algorithm optimizes the hyperparameters of CNN-BiLSTM, including the number of neurons in the convolutional layer, the size of the convolutional kernel, and the number of neurons in the long short-term memory layer.

[0076] Furthermore, in step six, the CNN-BiLSTM model includes four convolutional layers, three pooling layers, one flattening layer, one long short-term memory layer, and one output layer. The activation function of the convolutional layers is the ReLU function, with a stride of (2, 2) and zero padding. The pooling layer size is 2×2. The output layer outputs 4 cells, and the activation function is the softmax function. The number of neurons in the convolutional layers, the size of the convolutional kernels, and the number of neurons in the long short-term memory layer are obtained in step five. The method of this invention will be illustrated below using a specific rotating machinery fault case as an example.

[0077] The experimental data comes from the public dataset of the CWRU Bearing Data Center, and the experimental data information is shown in Table 1.

[0078] Table 1 Experimental Data Information Table

[0079]

[0080] The specific implementation process is as follows:

[0081] like Figure 1 A novel intelligent method for diagnosing faults in rotating machinery includes the following steps:

[0082] Step 1: Obtain the vibration signals of the rotating machinery under normal conditions and different fault types. In this embodiment, 70% of the total number of samples are selected as training samples, 10% as verification samples, and 20% as test samples.

[0083] Step 2: Variational Mode Decomposition (VMD) is used to preprocess the vibration signal for noise reduction, decomposing, denoising, and reconstructing the signal to generate a new vibration signal;

[0084] Step 3: Convert the new vibration signal into a two-dimensional time-frequency image using a short-time Fourier transform (STFT);

[0085] Step 4: Set the parameters required for the Sahara Silver Ant optimization algorithm, including the total number of individuals. M =30, the largest team size Ts max =5, minimum size Ts min =3, maximum number of iterations T =350;

[0086] Step 5: Optimize the hyperparameters of CNN-BiLSTM based on the Sahara Silver Ant optimization algorithm. The flowchart of the Sahara Silver Ant optimization algorithm is as follows: Figure 2 As shown, the optimized evaluation results are as follows: Figure 4 ;

[0087] Step Six: Input the two-dimensional time-frequency image into the optimized hyperparameter CNN-BiLSTM model to achieve rotating machinery fault diagnosis; the CNN-BiLSTM network model structure topology diagram is shown below. Figure 3 As shown, the evaluation results are shown in Table 2. Figure 5 , Figure 6 .

[0088] Step 7: Denoise the fault diagnosis signal to be diagnosed, convert it into a two-dimensional time-frequency image through short-time Fourier transform, and input the image into the trained rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM to realize rotating machinery fault diagnosis.

[0089] Table 2 Comparison of Model Accuracy and Loss Value

[0090]

Claims

1. A novel intelligent rotating machinery fault diagnosis method, characterized in that, The method comprises the following steps: Step one: obtaining vibration signals to be trained and fault classification of rotating machinery under normal conditions and different fault types; Step two: adopting variational mode decomposition to pre-process the vibration signals to be trained to obtain a reconstructed signal after pre-processing; Step three: converting the reconstructed signal into a two-dimensional time-frequency image through short-time Fourier transform; Step four: setting parameters required by the Sahara silver ant optimization algorithm; The Sahara silver ant optimization algorithm comprises five steps executed in sequence: Step S41: In the initialization phase, the Latin hypercube sampling method is used to generate M The initial distribution of the silver ants in the hot domain is mathematically expressed as follows: X k = lb +1 / M ×( Y + μ- 1)( ub - lb ) wherein, X k denotes the k position of the only silver ant individual; lb and ub denote the lower and upper bound of the search space, respectively, Y is a random permutation of {1,2,… M}, μ denotes a random number taken from the interval (0,1) ; Step S42: During the thermal adaptation and formation phase, silver ants randomly form teams to update their positions, with the number of members in each team... Ts It is dynamically adjusted, as follows: Ts = max(3, min(5, round(3 + 2sin(2π x 0.5))) t / T )))) wherein the round() function represents rounding the value inside the parentheses to an integer, t is the current iteration, T is the total number of iterations; Step S43: In the leader election stage, by evaluating each team in Ts The fitness value of the silver ant, and the individual with the optimal fitness value is selected as the leader of the team; Step S44: In the ant trail optimization stage, for each team leader, according to its fitness F i,b t between the current global optimal solution fitness F g t determines the search strategy: if F i,b t β F g t , path development is carried out in the direction towards the global optimum, where β is a behavior control parameter; otherwise, local exploration is directly carried out near the global optimal position, as follows:​​ β =1+1 / 2× U (0,1) X i,b t+1 = X i,b t +| N (0,1)|×( X g t - X i,b t )+ N (0,1)× τ ( n )× PF i,b t < β × F g t X i,b t+1 = X g t ×( U (0,1) +τ ( n ) ×P ) Other in, U (0,1) represents a uniformly distributed random variable in the interval [0,1]. F i,b t Representing the t In the nth iteration i The fitness score of each team leader F g t Representing the t The global optimal solution in the next iteration; X i,b t+1 and X i,b t They represent the first t+ 1st iteration and the 1st iteration t In the nth iteration i The position of a team leader X g t Indicates the first t The global optimal position in the next iteration. F i,b t and F g t The first t The fitness value of the leader of the i-th team relative to the global optimal position in the next iteration. N (0,1) is a standard normally distributed random variable. τ ( n (to obey the degree of freedom) n A random variable with a T-distribution, wherein n =5+round(10×(1–( t / T The round() function rounds the value within the parentheses to the nearest integer. P=(X k t -lb) / (ub-lb) ,in X k t Indicates the first k Only one silver ant individual in the first t Position in the next iteration; Step S45: During the exploration phase in unknown areas, for each team member, based on their fitness level... F i,m t Compatibility with team leaders F i,b t The proportional relationship between them determines the exploration strategy: if F i,m t <β×F i,b t If the search is successful, a cross-team learning approach is adopted, incorporating information from the best individuals in other teams for the search; otherwise, local exploration is performed directly near the current team leader. The update formula for this stage is as follows: X i,m t+1 =rand ( X i,b t ) + | N (0,1)| × ( rand ( X i,b t ) -X i,m t ) F i,m t <β×F i,b t X i,m t+1 = X i,b t ×( U (0,1) +τ ( n ) ×P ) Other wherein, X i,m t+1 and X i,m t denote the position of the t +1th iteration and the t th iteration, respectively, of a member of the team i m rand ( X i,b t ) denotes the position of a randomly chosen leader from the team leaders of the t th iteration;​​ After the five stages are calculated, it is determined whether the iteration is completed, if the iteration is completed, the search is ended, and the optimal position is output; if the iteration is not completed, step S42 is returned and steps S42-S45 are repeatedly executed; Step five: optimizing hyperparameters of a convolutional neural network-bidirectional long short-term memory network based on the Sahara silver ant optimization algorithm; Step six: training a rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM based on the optimized hyperparameters, the reconstructed signal and the fault classification; Step seven: pre-processing a fault diagnosis signal to be diagnosed, converting the signal into a two-dimensional time-frequency image through short-time Fourier transform, and inputting the image into the rotating machinery fault diagnosis model based on SSAO-VMD-CNN-BiLSTM trained to realize rotating machinery fault diagnosis.

2. A novel intelligent rotating machinery fault diagnosis method according to claim 1, characterized in that, In step two, the pre-processing includes the following procedures: setting the number of intrinsic mode functions and a penalty factor required for variational mode decomposition pre-processing; The pre-processing signal is decomposed by using the variational mode decomposition algorithm to obtain a plurality of intrinsic mode function components; the intrinsic mode functions are arranged in descending order of center frequency, the correlation of each intrinsic mode function is calculated, and the intrinsic mode function with the lowest correlation is marked as a noise intrinsic mode function; all the noise intrinsic mode functions are removed, and the remaining intrinsic mode functions are added to obtain a reconstructed signal after denoising.

3. A novel intelligent rotating machinery fault diagnosis method according to claim 1, characterized in that, In step four, the parameters required by the Sahara silver ant optimization algorithm include total number of individuals M , maximum team size Ts max , minimum size Ts min , maximum number of iterations T .

4. The novel intelligent rotating machinery fault diagnosis method according to claim 1, characterized in that, In step five, the Sahara silver ant optimization algorithm optimizes hyperparameters of the CNN-BiLSTM, including the number of convolutional layer neurons, the size of the convolution kernel, and the number of long short-term memory layer neurons.

5. The novel intelligent rotating machinery fault diagnosis method according to claim 1, characterized in that, In step six, the CNN-BiLSTM model comprises four convolutional layers, three pooling layers, one flattening layer, one long short-term memory layer, and one output layer, the activation function of the convolutional layer is a relu function, the stride is (2, 2), and zero padding is used; the size of the pooling layer is 2x2; the output of the output layer is 4, and the activation function is a softmax function; wherein the number of convolutional layer neurons, the size of the convolution kernel, and the number of long short-term memory layer neurons are obtained in step five.

Images