Bearing wear degree diagnosis method, device, medium, and product

Abstract

The application provides a bearing wear degree diagnosis method, device, medium and product, the method comprises the following steps: decomposing an acceleration vibration signal based on a variational mode decomposition method to obtain K sub-signals of different vibration modes, the acceleration vibration signal is an acceleration vibration signal of a target bearing; extracting feature information of each sub-signal of the different vibration modes based on a singular value decomposition method to obtain a feature component corresponding to each sub-signal of the different vibration modes; and sending the feature component to a gated recurrent unit to diagnose the wear degree of the target bearing through the gated recurrent unit. Therefore, the application improves the precision of bearing wear degree diagnosis.

Description

Technical Field

[0001] The present application relates to the technical field of bearing wear degree diagnosis, and in particular to a bearing wear degree diagnosis method, equipment, medium and product. Background Art

[0002] Armored vehicles are indispensable on the battlefield, and diesel engines are a crucial component in providing power. Their performance critically affects their effectiveness on the battlefield. Effectively monitoring the specific condition of diesel engines plays a vital role in protecting the safety of occupants and facilitating vehicle maintenance. Traditional methods for monitoring the wear of crankshaft bearings, a crucial component in diesel engines, often require significant manpower and resources, and suffer from low efficiency and accuracy. To address these shortcomings, leveraging modern information technology for mechanical fault diagnosis has become a hot topic among researchers worldwide.

[0003] Currently, new methods and theories are emerging in the field of bearing fault diagnosis, and many have been widely applied. Decomposing vibration signals based on the short-time Fourier transform has limitations and cannot effectively analyze non-stationary signals. Processing vibration signals using the wavelet transform lacks adaptability, and the selection of wavelet basis functions is crucial. Decomposing vibration signals using empirical mode decomposition can lead to modal aliasing. Decomposing vibration signals based on ensemble empirical mode decomposition cannot completely eliminate the effects of white noise, leading to reconstruction errors and other issues. Summary of the Invention

[0004] The present application provides a bearing wear degree diagnosis method, equipment, medium and product to solve the problems in the above-mentioned background technology.

[0005] In a first aspect, the present application provides a method for diagnosing bearing wear, comprising:

[0006] Decomposing the acceleration vibration signal based on the variational mode decomposition method to obtain K sub-signals of different vibration modes, wherein the acceleration vibration signal is the acceleration vibration signal of the target bearing;

[0007] Extracting characteristic information of each sub-signal of the different vibration modes based on a singular value decomposition method to obtain characteristic components corresponding to each sub-signal of the different vibration modes;

[0008] The characteristic components are sent to a gated cycle unit to diagnose the wear degree of the target bearing through the gated cycle unit.

[0009] Optionally, before decomposing the acceleration vibration signal based on the variational mode decomposition method, the method further includes:

[0010] The acceleration vibration signal is denoised based on a wavelet soft threshold method to obtain a denoised acceleration vibration signal.

[0011] Optionally, before decomposing the acceleration vibration signal based on the variational mode decomposition method, the method further includes:

[0012] Based on the first optimization algorithm, the penalty factor and the value of K are optimized.

[0013] Optionally, the first optimization algorithm is a tuna swarm optimization algorithm, a whale optimization algorithm or an ant colony optimization algorithm.

[0014] Optionally, before sending the feature component to the gated recurrent unit, the method further includes:

[0015] The gated recurrent unit is optimized based on a second optimization algorithm to obtain an optimized gated recurrent unit.

[0016] Optionally, the second optimization algorithm is a gold rush optimization algorithm, a whale optimization algorithm, a particle swarm optimization algorithm, a gray wolf optimization algorithm or a cuckoo optimization algorithm.

[0017] Optionally, the soft threshold denoising method is an unbiased likelihood estimation threshold method.

[0018] In a second aspect, an embodiment of the present application provides a computer electronic device, comprising:

[0019] One or more processors;

[0020] A memory for storing one or more instructions;

[0021] When one or more instructions are executed by one or more processors, the one or more processors are enabled to execute the method for implementing the above-mentioned first aspect.

[0022] In a third aspect, an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored, and the computer program is used to implement the method of the first aspect above.

[0023] In a fourth aspect, an embodiment of the present application provides a computer program product, including a computer program, which implements the method of the first aspect when executed by a processor.

[0024] The bearing wear degree diagnosis method provided by the embodiment of the application first decomposes the acceleration vibration signal of the bearing through VMD, VMD has good noise resistance and can be applied to non-stationary signals and can well analyze the acceleration vibration signal; then SVD is used to extract features from the decomposed vibration signal, which can not only effectively extract features but also simplify the processing amount and improve the processing efficiency; finally, GRO is used for wear degree diagnosis. The mode aliasing problem caused by the prior art of decomposing the acceleration vibration signal through empirical mode decomposition leads to low wear degree diagnosis precision. Therefore, compared with the prior art, the application improves the bearing wear degree diagnosis precision. Further, the application also uses SVD for decomposition in combination with VMD, which can not only effectively extract the features of the acceleration vibration signal but also simplify the processing amount and improve the wear degree diagnosis efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0026] Figure 1 An implementation environment architecture diagram of a bearing wear degree diagnosis method provided by the embodiment of the application;

[0027] Figure 2 A flowchart of a bearing wear degree diagnosis method provided by the embodiment of the application;

[0028] Figure 3 A flowchart of a bearing wear degree diagnosis method provided by the embodiment of the application;

[0029] Figure 4 A flowchart of another bearing wear degree diagnosis method provided by the embodiment of the application;

[0030] Figure 5 A flowchart of another bearing wear degree diagnosis method provided by the embodiment of the application;

[0031] Figure 6 A flowchart of a method for optimizing parameters K and penalty factors based on a tuna swarm optimization algorithm provided by the embodiment of the application;

[0032] Figure 7 A flowchart of a method for optimizing parameters of a gated recurrent unit based on a gold panning heat optimization algorithm provided by the embodiment of the application;

[0033] Figure 8It is a structural diagram of a computer system according to an embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application are clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work also fall within the scope of protection of this application. In addition, it should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0035] Figure 1 This is a diagram of an implementation environment architecture of a bearing wear degree diagnosis method according to an embodiment of the present application. Figure 1 As shown in FIG, the implementation environment architecture includes: sensors, terminals and servers.

[0036] Among them, the sensor is generally an industrial acceleration sensor, which is installed on the target bearing to be diagnosed and is used to collect the acceleration vibration signal of the target bearing.

[0037] The hardware structure of the terminal includes a processor, a memory and a display device, and the software structure includes a program product installed on the terminal.

[0038] The aforementioned program product may be independent software, such as an app deployed on a terminal as a client (e.g., a bearing wear diagnosis client). Of course, the aforementioned program product may also be a component of the terminal's operating system or an application, or a mini-program, for example, a mini-program in WeChat or an H5 (HTML5) webpage. It should be noted that a "mini-program" is an application that can be used without downloading or installing.

[0039] The above-mentioned program product can call the hardware on the terminal (such as one or more display screens, communication modules, etc.) to implement the technical solutions provided in this application.

[0040] Taking the program product as an application client deployed on the terminal as an example, when the application client is started, an application interface can be displayed on the terminal display screen, through which the parameters input by the user can be obtained, such as the value range of the number K of modal components mentioned later, and the input parameters can be sent to the server by calling the communication module, or one or more messages sent by the server can be received and displayed on the terminal display screen.

[0041] The types of terminals include but are not limited to smart phones, tablet computers, televisions, laptop computers, desktop computers, etc., and are not specifically limited in this embodiment of the present application.

[0042] The server is equipped with a variational mode decomposition model, a singular value feature extraction model, and a gated recurrent unit model. The server receives parameters sent by the terminal and uses these models to diagnose the degree of bearing wear. Furthermore, once the degree of wear is determined, the server sends the determined degree of wear to the terminal, which then displays the information on a display, allowing the user to understand the degree of bearing wear.

[0043] The server can be a single server, a server cluster consisting of multiple servers, or a cloud computing service center.

[0044] The terminal establishes a communication connection with the server through a wired or wireless network.

[0045] In addition to the aforementioned terminal + server model, the bearing wear diagnosis method can also be implemented by a separate electronic device, for example, a smartphone, tablet computer, television, laptop computer, desktop computer, etc. The variational mode decomposition model, singular value feature extraction model, and gated recurrent unit model can then be implemented in the separate electronic device.

[0046] Figure 2 This is a flow chart of a method for diagnosing bearing wear according to an embodiment of the present application. Figure 2 The method shown can be used by Figure 1 The server execution in Figure 2 As shown, combined with Figure 3 The method includes the following steps 201 to 203:

[0047] Step 201 : Decompose the acceleration vibration signal based on the variational mode decomposition method to obtain K sub-signals of different vibration modes. The acceleration vibration signal is the acceleration vibration signal of the target bearing.

[0048] Among them, variational mode decomposition (VMD) is used to decompose the acceleration vibration signal of the target bearing into K sub-signals of different vibration modes, that is, into K modal components. Each modal component has its own frequency center and bandwidth, so that step 202 can extract features from each modal component one by one, thereby improving the accuracy of feature extraction in step 202.

[0049] Wherein, K is the number of modal components, which is an integer greater than or equal to 2. Furthermore, the value of K can be determined by the first optimization algorithm mentioned below.

[0050] Each modal component is also called an intrinsic mode function (IMF) or an intrinsic mode function.

[0051] The acceleration vibration signal is collected by an industrial accelerometer and transmitted to the server. Furthermore, the industrial accelerometer is mounted at a specific location on the target bearing whose wear is to be diagnosed. Optionally, the industrial acceleration vibration sensor can be a PCB601A01 ICP industrial accelerometer.

[0052] Furthermore, the target bearing may be any bearing whose wear degree is desired to be detected, such as a crankshaft bearing of a diesel engine of a certain type of infantry fighting vehicle.

[0053] VMD can effectively analyze non-stationary signals and overcome the modal aliasing phenomenon that occurs in empirical mode decomposition (EMD) and ensemble empirical mode decomposition (EEMD), thereby improving the accuracy of decomposition and, in turn, the accuracy of bearing wear diagnosis.

[0054] Step 202 : extracting characteristic information of each sub-signal of the different vibration modes based on a singular value decomposition method to obtain characteristic components corresponding to each sub-signal of the different vibration modes.

[0055] Among them, singular value decomposition (SVD) has good stability and can simplify the processing of sub-signals of different vibration modes. Combined with VMD, it can not only extract features but also improve the efficiency of processing sub-signals of different vibration modes.

[0056] Furthermore, after extracting the characteristic information of each sub-signal of different vibration modes by singular value decomposition, characteristic components can be obtained. In this application, the sub-signals of different vibration modes after feature extraction by SVD are recorded as characteristic components.

[0057] For example, the first five sub-signals of the multiple different vibration modes in step 201 are combined into a matrix A. m×n , A=USV T , U is an m×m matrix, S is an m×n matrix, where S1=diag(σ1,σ2,…σ r ),σ1,σ2,……σ r are the singular values ​​of the matrix A, and σ1≥σ2≥,……≥σ r , r is the rank of matrix A. r The characteristic components are used as the feature set of the neural network. The rth characteristic component of the nth signal is σ nr , so the nth data is σ n2 , σ n2 ,...,σ n5 .

[0058] Or, in short, decompose the K sub-signals of different vibration modes obtained in step 201, take the first five sub-signals of different vibration modes to form a matrix, and calculate through known formulas to obtain the characteristic component, which is the characteristic of this sub-signal.

[0059] Since the acceleration vibration signal includes multiple signals of normal, light wear, moderate and heavy wear, the sub-signals of each different vibration mode are decomposed and extracted one by one to form a data set.

[0060] Step 203: Send the characteristic component to a gated circulation unit, so as to diagnose the wear degree of the target bearing through the gated circulation unit.

[0061] Among them, the gated recurrent unit network (GRU) can diagnose the degree of wear of the target bearing after training.

[0062] Compared with recurrent neural networks (RNNs), GRU neural networks can effectively capture semantic associations between long sequences and alleviate gradient explosion or vanishing phenomena. Compared with long short-term memory networks (LSTMs), GRU neural networks have a simpler internal structure and can effectively perform fault diagnosis.

[0063] Of course, the GRU needs to be trained and tested on the bearing's feature component set before use to obtain optimal parameters. The trained GRU is then used to diagnose the degree of bearing wear. Furthermore, the GRU's simple structure and low number of parameters significantly improve its training efficiency and reduce the risk of overfitting.

[0064] Furthermore, the update gate r of GRU t and reset gate z t The formula is as follows:

[0065] rt =σ(x t W xr +b r +W hr H t-1 ) (1)

[0066] z t =σ(x t W xz +b z +W hz H t-1 ) (2)

[0067] Among them, σ represents the sigmoid activation function, which takes a value between 0 and 1, and x t is the current input information, W and b represent the corresponding weight and bias values. t-1 is the hidden state at time t-1. t and z t Different values ​​will bring different results.

[0068] Candidate hidden states of GRU and hidden state H t

[0069]

[0070] Where tanh is the tanh activation function, and the symbol ⊙ refers to element-wise multiplication. Contains the hidden information entered previously, H t The calculation formula is the H that is forgotten to be passed down t-1 and adds some dimensional information of the current node input.

[0071] The wear degree may include normal, mild, moderate, and severe. For example, when the bearing wear is 0.05-0.11 mm, the target bearing is considered to be in a normal state; when the bearing wear is 0.11-0.15 mm, the target bearing is considered to be in a mild wear state; when the bearing wear is 0.15-0.20 mm, the target bearing is considered to be in a moderate wear state; and when the bearing wear is greater than 0.2 mm, the target bearing is considered to be in a severe wear state.

[0072] Alternatively, see Figure 4 and Figure 5 Before step 201, the method further includes step 200:

[0073] Step 200 , performing denoising processing on the acceleration vibration signal based on a wavelet soft threshold method to obtain a denoised acceleration vibration signal.

[0074] Furthermore, the soft threshold denoising method is an unbiased likelihood estimation threshold method.

[0075] The acceleration vibration signal x(t) collected by the sensor usually has external noise interference, so the obtained acceleration vibration signal x(t) is subjected to wavelet soft threshold denoising to obtain a new acceleration vibration signal x'(t).

[0076] In this case, the acceleration vibration signal mentioned in step 201 is the acceleration vibration signal after the noise reduction process in step 200 .

[0077] In addition, the number K of sub-signals of different vibration modes and the selection of the penalty factor α have a significant impact on the decomposition of the acceleration vibration signal. The specific reasons can be referred to the VMD process. Furthermore, the general process of VMD is as follows. For further introduction to VMD, please refer to the relevant existing technology.

[0078] The variational mode decomposition method decomposes the obtained signal x'(t) into K modal components u k (t), each IMF component has a finite bandwidth with a center frequency. The center frequency and bandwidth will be continuously updated during the decomposition process. Each IMF component is defined as an amplitude-frequency modulated signal, that is:

[0079] u k (t) = A k (t)cos(φ k (t)) (5)

[0080] Among them A k (t) is the amplitude of each modal component, φ k (t) is the instantaneous frequency of the signal.

[0081] In order to obtain the limited bandwidth of each modal component, it is necessary to k (t) performs Hilbert transform, and then u k (t) performs frequency shift, u k (t) frequency shift to ω k The constraint expression established is:

[0082]

[0083] in: represents the partial derivative with respect to t, δ(t) is the impulse function, {u k (t)}={u1,u2……,u k} are the K modal components obtained by decomposition, {ω k}={ω1,ω2,……,ω k} is the center frequency of each component.

[0084] In order to solve the above equation {u k (t)} and {ω k}, introducing the quadratic penalty factor α and the Lagrange multiplication operator λ, the constrained variational problem is transformed into an unconstrained variational problem:

[0085]

[0086] The multiplication operator alternating direction method is used to k (t)},{ω k} and λ are optimized iteratively, and the expression for iterative update is:

[0087]

[0088] Therefore, from the above process, it can be seen that the selection of modal component K and penalty factor α has a great influence on the decomposition. Therefore, in order to improve the decomposition accuracy of VMD, the optimization algorithm can be used to optimize the two parameters of modal component K and penalty factor α to achieve the purpose of parameter optimization.

[0089] Alternatively, see Figure 4 and Figure 5 Before executing step 201, the method further includes the following step 204:

[0090] Step 204: Optimize the K value and penalty factor based on the first optimization algorithm.

[0091] Optionally, the first optimization algorithm is a tuna swarm optimization algorithm (full name in English: Tuna Swarm Optimization, English abbreviation: TSO), a whale optimization algorithm (full name in English: Whale Optimization Algorithm; English abbreviation: WOA) or an ant colony optimization algorithm (full name in English: Ant Colony Optimization, English abbreviation: ACO).

[0092] In this case, the variational mode decomposition method used in step 201 is a variational mode decomposition method optimized by the first optimization algorithm.

[0093] Further, see Figure 6 This application uses the TSO optimization algorithm as an example to illustrate the process of the TSO optimization algorithm.

[0094] First, the fish school is initialized:

[0095]

[0096] in, is the i-th initial individual, ub and lb represent the upper and lower boundaries of the search space, N is the number of tuna schools, and rand is a random vector ranging from 0 to 1.

[0097] Secondly, schools of fish search for food in spirals. To find food, schools of tuna will form a spiral structure to chase food. The calculation formula is:

[0098]

[0099] α1 and α2 are obtained from the following equations (13) and (14), β is the spiral factor, and β = e bl ×cos(2πb), a and e are both constants.

[0100]

[0101] In addition to the aforementioned spiral foraging pattern, tuna schools also use parabolic foraging. Assuming the probability of selecting both methods is 50%, the calculation formula is as follows:

[0102]

[0103] The value of TF can only be a random number of 1 or -1, and p determines the development range of the tuna school.

[0104] Furthermore, the tuna school optimization algorithm is used to optimize the number of components and penalty factors in the variational mode decomposition. Furthermore, the range of k and α corresponds to the range within which the tuna school can find food, regardless of the number of tuna in the school. Each fish position in the tuna school represents a solution, and the optimal tuna position is the optimal solution, which is also the optimal solution for k and a. Therefore, the tuna school optimization algorithm is used to optimize K and α. The steps are as follows:

[0105] 1) Initialize the tuna school parameters, position vector [k,α];

[0106] 2) Perform VMD decomposition, calculate the fitness of each tuna and update the parameters, and select permutation entropy as the fitness function;

[0107] 3) Update the position of each tuna;

[0108] 4) If the number of iterations is met, output the parameters [k, α] and bring the parameters into VMD for operation; otherwise, return to step 2 and proceed to the next iteration; until the number of iterations is met, output the parameters [k, α].

[0109] After the optimization is completed, k and α can be brought into VMD, and the optimized VMD can be used to decompose the acceleration vibration signal of the target bearing, thereby improving the decomposition efficiency and accuracy of VMD.

[0110] Alternatively, see Figure 4 and Figure 5 Before step 203, the method further includes the following step 205:

[0111] Step 205 : Optimize the gated recurrent unit based on a second optimization algorithm to obtain an optimized gated recurrent unit.

[0112] That is, the gated recurrent unit in step 203 may be a gated recurrent unit optimized using the second optimization algorithm.

[0113] Optionally, the second optimization algorithm is a gold rush optimization algorithm (full name in English: Gold Rush Optimizer; English abbreviation: GRO), a whale optimization algorithm (WOA), a particle swarm optimization algorithm (full name in English: Particle Swarm Optimization; English abbreviation: PSO), a grey wolf optimization algorithm (full name in English: Grey Wolf Optimizer; English abbreviation: GWO), a cuckoo optimization algorithm (full name in English: Cuckoo Optimization Algorithm; English abbreviation: COA), etc.

[0114] For example, see Figure 7 , using the Gold Rush optimization algorithm as an example:

[0115] When constructing and training GRU, the parameters involved have a great impact on the accuracy of bearing wear diagnosis. To solve this problem, GRO is used to optimize the parameters of GRU.

[0116] First, randomly generate a group of gold diggers:

[0117]

[0118] F GP is the fitness value corresponding to each gold digger:

[0119]

[0120] Secondly, the prospector migration stage is carried out. The formula for simulating the migration of gold prospectors to gold mines is:

[0121]

[0122] Again, during the gold mining stage, each gold digger will mine more gold in the gold mining area.

[0123]

[0124] Then, there is the stage of cooperation among gold prospectors.

[0125]

[0126] Finally, the prospector relocates. To decide whether the prospector should remain in its original location or move to a new one, the two locations are compared using an evaluation function. During this process, if the value of the objective function improves, the prospector updates its location; otherwise, it remains in its previous location, as shown in the following equation:

[0127]

[0128] The gated recurrent unit is used as the objective function f() of the gold rush optimization algorithm, which mainly controls the initial learning rate, the number of hidden layer nodes and L 2 The regularization parameters are set as independent variables x1, x2, and x3 respectively.

[0129] It should be noted here that the gold prospector position represents a feasible solution. Finding the optimal prospector position is also finding the optimal solution. The optimal solution is the optimal parameters. The steps are as follows:

[0130] 1) Initialization: number of gold miners y, location of gold prospectors Iteration number t, gold mine location

[0131] 2) When any gold prospector migrates to a new location, the fitness of the current new location is calculated using known related methods.

[0132] 3) At the current time, determine the position of the gold prospector with the highest fitness and obtain the position of the best gold prospector.

[0133] 4) Each gold prospector moves to a new location using one of the migration methods defined in the GRO algorithm (gold mining or collaboration).

[0134] 5) Update the best solution (gold mine).

[0135] 6) Repeat steps 2 to 5 until the number of iterations is met, and then we get [1].

[0136] 7) Output the optimal parameters to the GRU network.

[0137] At this point, the GRU network with optimal parameters can be used to diagnose the degree of bearing wear.

[0138] The GRO-optimized GRU neural network model has higher prediction accuracy and higher practicability than the original GRU. The diagnostic accuracy of the unoptimized gated recurrent unit network is only 91.6667%, and although the accuracy is above 90%, it fails to meet the 95% standard, while the accuracy of the GRO-optimized gated recurrent unit network for diagnosing the wear degree of the diesel engine crankshaft bearing is as high as 97.9167%, and the optimization effect is obvious. This model can effectively diagnose the wear degree of the target bearing based on the obtained vibration signals.

[0139] The bearing wear degree diagnosis method provided by the embodiments of the present application first decomposes the acceleration vibration signal of the bearing through VMD. VMD has good noise resistance and can be applied to non-stationary signals and can well analyze the acceleration vibration signal. Then, SVD is used to extract features from the decomposed vibration signal, which can effectively extract features, simplify the processing amount and improve the processing efficiency. Finally, GRO is used for wear degree diagnosis. The mode aliasing problem caused by the decomposition of the acceleration vibration signal through the empirical mode decomposition in the prior art leads to low wear degree diagnosis accuracy. Therefore, compared with the prior art, the present application improves the bearing wear degree diagnosis accuracy. Further, the present application also uses SVD for decomposition in combination with VMD, which can effectively extract the features of the acceleration vibration signal, simplify the processing amount and improve the wear degree diagnosis efficiency.

[0140] Figure 8 Fig. 1 is a structural schematic diagram of an electronic device or computer system 800 according to an embodiment of the present application. The computer system includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 802 or programs loaded from a storage portion into a random access memory (RAM) 803. In the RAM 803, various programs and data required for system operation are also stored. The CPU 801, the ROM 802, and the RAM 803 are connected to each other through a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0141] The following components are connected to the I / O interface 805: an input portion 806 including a keyboard, a mouse, and the like; an output portion 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), a speaker, and the like; a storage portion 808 including a hard disk and the like; and a communication portion 809 including a network interface card such as a LAN card, a modem, and the like. The communication portion 809 performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface 805 as necessary. A removable medium 811 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 810 as necessary, so that a computer program read therefrom is installed in the storage portion 808 as necessary.

[0142] In particular, the processes described in the flowcharts of the embodiments of the present application can be implemented as computer software programs. For example, the method embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart.

[0143] As another aspect, the present application further provides a computer-readable medium, which may be included in the electronic device described in the above embodiments, or may exist independently without being incorporated into the electronic device. The computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device implements the method described in the embodiments of the present application.

[0144] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two.

[0145] The units involved in the embodiments described in this application can be implemented by software or hardware, and the units described can also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves. The units or modules described can also be set in a processor. In some cases, the names of these units or modules do not constitute limitations on the units or modules themselves.

[0146] In addition, it should be noted that the scope of this application includes feasible technical solutions formed by specific combinations of the above-mentioned technical features, and should also cover other feasible technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned application concept.

[0147] Finally, it should be noted that any content not described in the technical solution of this application can be implemented using existing technologies. In addition, the above embodiments are only used to illustrate the technical solution of this application, and not to limit it. Although this application has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some or all of the technical features therein can be replaced with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solution from the scope of the technical solution of the embodiments of this application.

Claims

1. A method for diagnosing bearing wear, characterized in that: include: Decomposing the acceleration vibration signal based on the variational mode decomposition method to obtain K sub-signals of different vibration modes, wherein the acceleration vibration signal is the acceleration vibration signal of the target bearing; Extracting characteristic information of each sub-signal of the different vibration modes based on a singular value decomposition method to obtain characteristic components corresponding to each sub-signal of the different vibration modes; Sending the characteristic component corresponding to each of the sub-signals of the different vibration modes to a gated circulation unit, so as to diagnose the wear degree of the target bearing through the gated circulation unit, where the wear degree includes normal, mild, moderate and severe; Before decomposing the acceleration vibration signal based on the variational mode decomposition method, the method further includes: Optimizing a penalty factor of the variational mode decomposition method and a value of K based on a first optimization algorithm, wherein the penalty factor and the value of K affect the accuracy of the variational mode decomposition method; The first optimization algorithm is a tuna swarm optimization algorithm, a whale optimization algorithm or an ant colony optimization algorithm.

2. The method for diagnosing bearing wear according to claim 1, characterized in that: Before decomposing the acceleration vibration signal based on the variational mode decomposition method, the method further includes: The acceleration vibration signal is denoised based on a wavelet soft threshold method to obtain a denoised acceleration vibration signal.

3. The method for diagnosing bearing wear according to claim 1, characterized in that: Before sending the feature component to the gated recurrent unit, the method further includes: The gated recurrent unit is optimized based on a second optimization algorithm to obtain an optimized gated recurrent unit.

4. The method for diagnosing bearing wear according to claim 3, characterized in that: The second optimization algorithm is a gold rush optimization algorithm, a whale optimization algorithm, a particle swarm optimization algorithm, a gray wolf optimization algorithm or a cuckoo optimization algorithm.

5. The method for diagnosing bearing wear according to claim 2, characterized in that: The soft threshold denoising method is an unbiased likelihood estimation threshold method.

6. A computer electronic device, characterized in that: The device comprises: A memory for storing one or more instructions; one or more processors, configured to execute the one or more instructions; When the one or more instructions are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that A computer program is stored thereon, the computer program being configured to: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

8. A computer program product, characterized in that The method comprises a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images