Methods, devices, equipment, and readable storage media for monitoring bearing health status

Abstract

This invention discloses a method, apparatus, equipment, and readable storage medium for monitoring the health status of bearings. It acquires multi-source information data of the bearing during the monitoring phase; processes the multi-source information data based on a trained bearing performance degradation monitoring model to obtain the bearing's performance degradation index during the monitoring phase. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using multi-source information data of the bearing during the initial operation phase; and determines the bearing's health status based on the deviation between the bearing's performance degradation index during the monitoring phase and a pre-acquired performance degradation index during the initial operation phase. This invention improves the accuracy of bearing health status monitoring and enables real-time monitoring of bearing health status.

Description

Technical Field

[0001] This invention belongs to the field of mechanical equipment performance degradation assessment, and specifically relates to a method, device, equipment and readable storage medium for monitoring the health status of bearings. Background Technology

[0002] Under the general trend of intelligent development in industrial equipment, real-time monitoring of equipment operating status and degradation indicators is crucial to ensuring the continuous healthy operation of equipment. Bearings, as critical rotating components widely found in mechanical equipment, are constantly subjected to high-temperature, high-speed, and complex environments under certain loads; their health status has a significant impact on the operating status of industrial equipment. Therefore, real-time acquisition of potential bearing degradation information and monitoring of bearing operation and degradation status are the foundation and prerequisite for healthy maintenance strategies for mechanical equipment.

[0003] Currently, most scholars, both domestically and internationally, utilize traditional machine learning methods for bearing degradation monitoring. However, this inevitably requires complex manual preprocessing of equipment signals, and most studies employ single-source data. This leads to low accuracy when monitoring time-series data with complex multi-source information distributions and noise. Furthermore, most existing algorithms require the use of complex evaluation metrics to screen proposed features. This process is not only time-consuming and labor-intensive, but it is also difficult to determine whether the selected features are truly beneficial to the final degradation assessment. Currently, there is no effective method to characterize the bearing degradation state under multi-source information data for real-time health monitoring. Summary of the Invention

[0004] To address the problems existing in the prior art, the present invention provides a method, apparatus, device, and readable storage medium for monitoring the health status of bearings, thereby improving the accuracy of bearing health status monitoring and enabling real-time monitoring of bearing health status.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0006] A method for monitoring the health status of a bearing, comprising:

[0007] Acquire multi-source information data of bearings during the monitoring phase;

[0008] The multi-source information data of the monitoring stage is processed based on the trained bearing performance degradation monitoring model to obtain the performance degradation index of the bearing in the monitoring stage. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using the multi-source information data of the bearing in the initial operation stage.

[0009] The health status of the bearing is determined based on the degree of deviation between the bearing's performance degradation indicators during the monitoring phase and the bearing's performance degradation indicators during the initial operation phase, which were obtained in advance.

[0010] Furthermore, the step of training an adversarial fusion convolutional autoencoder using multi-source information data from the bearing during its initial operation phase includes:

[0011] The multi-source information data of the bearing during the initial operation phase is obtained as training samples.

[0012] The training samples are input into the adversarial fusion convolutional autoencoder, which includes an encoder unsupervised fusion feature extraction module, a Gaussian distribution random sampling module, a decoder signal reconstruction module, and a computation module, wherein:

[0013] The encoder unsupervised fusion feature extraction module encodes the multi-source information data of the bearing during the initial operation phase to obtain the unsupervised fusion features of the multi-source information data of the bearing during the initial operation phase, and maps the unsupervised fusion features into latent variables in the latent space.

[0014] The Gaussian distribution random sampling module maps the probability distribution of the latent variables to a Gaussian distribution, thereby obtaining latent variables that follow a Gaussian distribution;

[0015] The decoder signal reconstruction module decodes the latent variables that follow a Gaussian distribution to obtain the reconstructed multi-source information data of the bearing during the initial operation phase.

[0016] The calculation module calculates the mean square error between the reconstructed bearing's multi-source information data during the initial operation phase and the bearing's multi-source information data during the initial operation phase, thus completing the training of the bearing performance degradation monitoring model.

[0017] Furthermore, the anti-fusion convolutional autoencoder also includes a discriminant network adversarial learning module, which performs adversarial training on the reconstructed multi-source information data of the bearing during the initial operation phase and the multi-source information data of the bearing during the initial operation phase.

[0018] Furthermore, the calculation of the mean square error between the reconstructed multi-source information data of the bearing during the initial operating phase and the multi-source information data of the bearing during the initial operating phase specifically involves:

[0019]

[0020] In the formula: N is the size of the sample batch, x i This refers to the multi-source information data of the bearing during its initial operation phase. This refers to the multi-source information data of the reconstructed bearing during the initial operation phase.

[0021] Furthermore, the method for obtaining the performance degradation index of the bearing during the initial operation phase is as follows:

[0022] Acquire multi-source information data of the bearing during the initial operation phase;

[0023] The multi-source information data of the bearing during the initial operation phase is input into the bearing performance degradation monitoring model, and the performance degradation index of the bearing during the initial operation phase is output.

[0024] Furthermore, the health status of the bearing is determined based on the degree of deviation between the bearing's performance degradation indicators during the monitoring phase and the pre-acquired performance degradation indicators during the initial operation phase, including:

[0025] The greater the deviation between the performance degradation index of the bearing during the monitoring phase and the pre-obtained performance degradation index of the bearing during the initial operation phase, the worse the health condition of the bearing.

[0026] A bearing health status monitoring device, comprising:

[0027] The acquisition module is used to acquire multi-source information data of the bearing during the monitoring phase;

[0028] The processing module is used to process the multi-source information data of the monitoring stage based on the trained bearing performance degradation monitoring model to obtain the performance degradation index of the bearing in the monitoring stage. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using the multi-source information data of the bearing in the initial operation stage.

[0029] The judgment module is used to determine the health status of the bearing based on the degree of deviation between the performance degradation index of the bearing during the monitoring phase and the performance degradation index of the bearing during the initial operation phase obtained in advance.

[0030] An apparatus includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for monitoring the health status of a bearing.

[0031] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for monitoring the health status of a bearing.

[0032] Compared with the prior art, the present invention has at least the following beneficial effects:

[0033] The bearing health status monitoring method provided by this invention acquires multi-source information data of the bearing during the monitoring phase and processes this data based on a trained bearing performance degradation monitoring model to obtain bearing performance degradation indices during the monitoring phase. The bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using multi-source information data of the bearing during its initial operation phase. The bearing health status is determined based on the deviation between the bearing performance degradation indices during the monitoring phase and pre-acquired performance degradation indices during the initial operation phase. This invention effectively solves the problem of bearing health status monitoring under multi-source information data. The encoder unsupervised fusion feature extraction module in the constructed adversarial fusion convolutional autoencoder can directly extract the fusion features of multi-source information data without requiring manual feature selection based on monitoring results. Furthermore, it adapts to the complex distribution of multi-source input data and effectively addresses the problem of low accuracy in bearing operational health status monitoring under single-source information data.

[0034] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the specific embodiments of the present invention, the drawings used in the description of the specific embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0036] Figure 1 This is a flowchart of a bearing health status monitoring method based on multi-source information according to the present invention.

[0037] Figure 2 This is a diagram of the adversarial fusion convolutional autoencoder structure proposed in this invention;

[0038] Figure 3 This is a graph of vibration signal data of the bearing throughout its entire lifespan (horizontal direction).

[0039] Figure 4 This is a graph of vibration signal data (vertical direction) of the bearing throughout its entire lifespan.

[0040] Figure 5 Training graph for reconfiguration loss of bearings throughout their entire lifespan;

[0041] Figure 6 This is a graph showing the remodeling loss and degradation index of a bearing throughout its entire lifespan. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. 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.

[0043] As a specific embodiment of the present invention, combined with Figure 1 As shown, a method for monitoring the health status of a bearing specifically includes the following steps:

[0044] Step 1: Obtain multi-source information data of the bearing during the monitoring phase.

[0045] Step 2: Process the multi-source information data of the monitoring stage based on the trained bearing performance degradation monitoring model to obtain the performance degradation index of the bearing in the monitoring stage. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using the multi-source information data of the bearing in the initial operation stage.

[0046] Specifically, training an adversarial fusion convolutional autoencoder using multi-source information data from the bearing during its initial operation phase includes:

[0047] The multi-source information data of the bearing during the initial operation phase is obtained as training samples.

[0048] The training samples are input into the adversarial fusion convolutional autoencoder, which includes an encoder unsupervised fusion feature extraction module, a Gaussian distribution random sampling module, a decoder signal reconstruction module, a discriminant network adversarial learning module, and a computation module, wherein:

[0049] The encoder unsupervised fusion feature extraction module encodes the multi-source information data of the bearing during the initial operation phase to obtain the unsupervised fusion features of the multi-source information data of the bearing during the initial operation phase, and maps the unsupervised fusion features into latent variables in the latent space.

[0050] The Gaussian distribution random sampling module maps the probability distribution of the latent variables to a Gaussian distribution, thus obtaining latent variables that follow a Gaussian distribution.

[0051] The decoder signal reconstruction module decodes the latent variables that follow a Gaussian distribution to obtain the reconstructed multi-source information data of the bearing during the initial operation phase.

[0052] The discriminative network adversarial learning module performs adversarial training on the reconstructed multi-source information data of the bearing during the initial operation phase and the multi-source information data of the bearing during the initial operation phase, thereby enhancing the network regularization effect and improving the sensitivity of health indicators.

[0053] The calculation module calculates the mean square error between the reconstructed bearing's multi-source information data during the initial operation phase and the bearing's multi-source information data during the initial operation phase, thus completing the training of the bearing performance degradation monitoring model.

[0054] The calculation of the mean square error between the reconstructed bearing's multi-source information data during the initial operating phase and the bearing's multi-source information data during the initial operating phase is specifically as follows:

[0055]

[0056] In the formula: N is the size of the sample batch, x i This refers to the multi-source information data of the bearing during its initial operation phase. This refers to the multi-source information data of the reconstructed bearing during the initial operation phase.

[0057] Step 3: Determine the health status of the bearing based on the degree of deviation between the performance degradation index of the bearing during the monitoring phase and the performance degradation index of the bearing during the initial operation phase obtained in advance.

[0058] Specifically, the method for obtaining the performance degradation index of the bearing during the initial operation phase is as follows:

[0059] Acquire multi-source information data of the bearing during the initial operation phase;

[0060] The multi-source information data of the bearing during the initial operation phase is input into the bearing performance degradation monitoring model, and the performance degradation index of the bearing during the initial operation phase is output.

[0061] The following detailed description uses a set of accelerated fatigue life tests on bearings as verification of the present invention.

[0062] a. The bearing speed was set to 2100 r / min, and a radial force of 12 kN was applied. Vibration signals from the bearing housing were collected using an accelerometer at a sampling frequency of 25.6 kHz, with sampling occurring every 1 minute for 1.28 seconds. Vibration signals from two channels, one horizontal and one vertical, were collected, thus completing the acquisition of multi-source bearing data. Figure 3 and Figure 4 As shown.

[0063] b. Figure 3As shown, the vibration amplitude of the first 20 sets of samples collected during the initial operation phase of the bearing (i.e., the bearing health phase) is stable and basically remains at a low level. Therefore, the first 20 sets of multi-source information data are used as training samples for training the bearing performance degradation monitoring model.

[0064] c. Use the first 20 sets of multi-source information data of the bearing as input to train the adversarial fusion convolutional autoencoder, such as... Figure 2 As shown, the adversarial fusion convolutional autoencoder includes an encoder unsupervised fusion feature extraction module, a Gaussian distribution random sampling module, a decoder signal reconstruction module, a discriminant network adversarial learning module, and a computation module, specifically:

[0065] The first 20 sets of multi-source information data of the bearing are used as the input of the adversarial fusion convolutional autoencoder. The encoder's unsupervised fusion feature extraction module encodes the input signal and outputs the unsupervised fusion features of the first 20 sets of multi-source information data. The unsupervised fusion features are mapped as latent variables in the latent space.

[0066] The Gaussian distribution random sampling module maps the probability distribution of the latent variables to a Gaussian distribution, obtaining latent variables that follow a Gaussian distribution. Since the input data in this study is a one-dimensional time-series signal, the probability density function of the random sampling is as follows:

[0067]

[0068] Where: z is a random variable, μ is the expected value, and σ is the expected value. 2 Let Variance be the variance.

[0069] The sampling discriminative network adversarial learning model training strategy gradually fits the distribution of the encoder-generated data to a Gaussian distribution, thereby achieving the regularization effect of the entire autoencoder.

[0070] The decoder signal reconstruction module decodes the latent variables that follow a Gaussian distribution to obtain the reconstructed multi-source information data of the bearing during the initial operation phase. Let x... i This is the input data for the encoder's unsupervised fusion feature extraction module. The reconstructed data output by the decoder signal reconstruction module is used to calculate the bearing degradation index after obtaining the reconstructed data, as shown in the following formula:

[0071]

[0072] In the formula: N is the size of the sample batch, x i This refers to the multi-source information data of the bearing during its initial operation phase. This refers to the multi-source information data of the reconstructed bearing during the initial operation phase.

[0073] After training, the reconstruction loss of the multi-source information data of the bearing during the initial operation phase converged to approximately 0.35. Figure 5 As shown, 0.35 is therefore set as the bearing degradation index during the initial operation phase.

[0074] d. Continue to acquire vibration signals of the bearing in both horizontal and vertical directions during the monitoring phase, i.e., multi-source information data.

[0075] e. Input the multi-source information data of the bearing during the monitoring phase into the trained bearing performance degradation monitoring model, and output the reconstruction loss of the multi-source information data of the bearing during the monitoring phase, that is, the performance degradation index of the bearing during the monitoring phase.

[0076] f. such as Figure 6 As shown, the performance degradation index of the bearing remained basically unchanged before the 30th sample, stabilizing around the reconstructed value of 0.35. However, between the 30th and 40th samples, the performance degradation index began to fluctuate to some extent, indicating that the bearing began to fail and entered the initial stage of bearing degradation. As the degree of failure deepened, the bearing performance degradation index value gradually deviated from the threshold. The higher the degree of deviation, the more severe the bearing failure, thus completing the real-time degradation monitoring of the bearing.

[0077] This invention provides a bearing health status monitoring system for implementing the monitoring method of this invention, comprising:

[0078] The acquisition module is used to acquire multi-source information data of the bearing during the monitoring phase;

[0079] The processing module is used to process the multi-source information data of the monitoring stage based on the trained bearing performance degradation monitoring model to obtain the performance degradation index of the bearing in the monitoring stage. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using the multi-source information data of the bearing in the initial operation stage.

[0080] The judgment module is used to determine the health status of the bearing based on the degree of deviation between the performance degradation index of the bearing during the monitoring phase and the performance degradation index of the bearing during the initial operation phase obtained in advance.

[0081] In one embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used to implement the operation of a method for monitoring the health status of bearings.

[0082] In one embodiment of the present invention, a method for monitoring the health status of a bearing, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data.

[0083] The computer storage medium can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs)), optical storage (e.g., CDs, DVDs, BDs, HVDs), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0084] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for monitoring the health status of a bearing, characterized in that, include: Acquire multi-source information data of bearings during the monitoring phase; The multi-source information data of the monitoring stage is processed based on the trained bearing performance degradation monitoring model to obtain the performance degradation index of the bearing in the monitoring stage. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using the multi-source information data of the bearing in the initial operation stage. The method of training an adversarial fusion convolutional autoencoder using multi-source information data from the bearing during its initial operation phase includes: The multi-source information data of the bearing during the initial operation phase is obtained as training samples. The training samples are input into the adversarial fusion convolutional autoencoder, which includes an encoder unsupervised fusion feature extraction module, a Gaussian distribution random sampling module, a decoder signal reconstruction module, a computation module, and a discriminant network adversarial learning module, wherein: The encoder unsupervised fusion feature extraction module encodes the multi-source information data of the bearing during the initial operation phase to obtain the unsupervised fusion features of the multi-source information data of the bearing during the initial operation phase, and maps the unsupervised fusion features into latent variables in the latent space. The Gaussian distribution random sampling module maps the probability distribution of the latent variables to a Gaussian distribution, thereby obtaining latent variables that follow a Gaussian distribution; The decoder signal reconstruction module decodes the latent variables that follow a Gaussian distribution to obtain the reconstructed multi-source information data of the bearing during the initial operation phase. The calculation module calculates the mean square error between the reconstructed multi-source information data of the bearing during the initial operation phase and the multi-source information data of the bearing during the initial operation phase. The discriminant network adversarial learning module performs adversarial training on the reconstructed multi-source information data of the bearing during the initial operation phase and the multi-source information data of the bearing during the initial operation phase. The health status of the bearing is determined based on the degree of deviation between the bearing's performance degradation indicators during the monitoring phase and the bearing's performance degradation indicators during the initial operation phase, which were obtained in advance.

2. The method for monitoring the health status of a bearing according to claim 1, characterized in that, The calculation of the mean square error between the reconstructed bearing's multi-source information data during the initial operating phase and the bearing's multi-source information data during the initial operating phase is specifically as follows: In the formula: This refers to the size of the sample batch. This refers to the multi-source information data of the bearing during its initial operation phase. This refers to the multi-source information data of the reconstructed bearing during the initial operation phase.

3. The method for monitoring the health status of a bearing according to claim 1, characterized in that, The method for obtaining the performance degradation index of the bearing during the initial operation phase is as follows: Acquire multi-source information data of the bearing during the initial operation phase; The multi-source information data of the bearing during the initial operation phase is input into the bearing performance degradation monitoring model, and the performance degradation index of the bearing during the initial operation phase is output.

4. The method for monitoring the health status of a bearing according to claim 1, characterized in that, The health status of the bearing is determined based on the degree of deviation between the bearing's performance degradation indicators during the monitoring phase and the pre-acquired performance degradation indicators during the initial operation phase. This includes: The greater the deviation between the performance degradation index of the bearing during the monitoring phase and the pre-obtained performance degradation index of the bearing during the initial operation phase, the worse the health condition of the bearing.

5. A device for monitoring the health status of a bearing, characterized in that, A method for monitoring the health status of a bearing as described in any one of claims 1 to 4, comprising: The acquisition module is used to acquire multi-source information data of the bearing during the monitoring phase; The processing module is used to process the multi-source information data of the monitoring stage based on the trained bearing performance degradation monitoring model to obtain the performance degradation index of the bearing in the monitoring stage. The trained bearing performance degradation monitoring model is obtained by training an adversarial fusion convolutional autoencoder using the multi-source information data of the bearing in the initial operation stage. The judgment module is used to determine the health status of the bearing based on the degree of deviation between the performance degradation index of the bearing during the monitoring phase and the performance degradation index of the bearing during the initial operation phase obtained in advance.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of a bearing health status monitoring method as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of a bearing health status monitoring method as described in any one of claims 1 to 4.

Images