Press machine bearing fault diagnosis method and system, terminal and medium

Abstract

The invention belongs to the technical field of bearing fault diagnosis, and particularly discloses a press machine bearing fault diagnosis method and system, a terminal and a medium, and the method comprises the steps: collecting multi-source signal data of a press machine bearing in an operation process; performing denoising and standardization processing on the signal, and extracting time domain and frequency domain features; different sensor features are fused, and multi-dimensional feature representation is constructed; inputting the features into a deep learning model combining a convolutional neural network and a long-short-term memory network for training and reasoning, and introducing a multi-kernel maximum mean value difference strategy to realize feature distribution alignment; and performing fault identification on the real-time signal based on a multi-label classification mode, and triggering an alarm when the prediction probability meets a preset condition. According to the method, fusion modeling of multi-source data and accurate recognition of composite faults are achieved, and the method has high cross-working-condition adaptive capacity and real-time diagnosis capacity and is suitable for intelligent operation and maintenance scenes of press equipment.

Description

Technical Field

[0001] The present invention belongs to the technical field of bearing fault diagnosis, and in particular relates to a method, system, terminal and medium for diagnosing a press bearing fault. Background Art

[0002] The bearings of presses on high-speed stamping production lines are crucial moving components in mechanical equipment such as presses, and their performance directly determines the stability and operational efficiency of the equipment. As presses operate for long periods of time and under high loads, the bearings are constantly subjected to wear and stress from the working environment, and may suffer varying degrees of damage, such as fatigue wear, cracks, and surface peeling. If these faults are not detected and repaired promptly, they can lead to bearing failure, which in turn can cause serious equipment failures and even affect the normal operation of the entire production line. Bearing failures not only increase repair costs but also extend downtime, resulting in a significant decrease in production efficiency and even endangering the safety of personnel and equipment. Therefore, accurately and in real time diagnosing bearing faults during press operation is key to ensuring efficient, stable, and safe production.

[0003] Traditional bearing fault diagnosis methods typically rely on the acquisition and analysis of vibration signals. Common approaches include feature extraction and analysis based on the time or frequency domain. These methods determine the health of the bearing by extracting signal features such as vibration amplitude, frequency content, and periodic variations. However, time and frequency domain feature analysis methods have inherent limitations, primarily in their inability to effectively capture the complex nonlinear and time-varying characteristics of bearing faults. Especially when bearing faults exhibit nonlinearity and nonstationarity, these traditional methods often struggle to fully and accurately reflect the actual bearing fault condition. Furthermore, traditional methods are highly sensitive to noise, particularly in complex operating conditions and high-noise environments, which can lead to decreased diagnostic accuracy and, in turn, impact the timeliness and reliability of fault detection. Summary of the Invention

[0004] In response to the problems in the prior art, the present invention provides a method, system, terminal and medium for diagnosing press bearing faults, which solves the problem that when the bearing fault is nonlinear and non-stationary, the traditional methods of the prior art are often difficult to fully and accurately reflect the actual fault condition of the bearing.

[0005] The technical solution adopted in the present invention is as follows: In a first aspect, the present application provides a method for diagnosing a press bearing fault, the method comprising the following steps: Step S1, collecting multi-source signal data of a press bearing during operation, where the multi-source signal data includes a vibration signal, a temperature signal, and a pressure signal; Step S2: preprocessing the multi-source signals, including signal denoising and standardization operations; Get the preprocessed time series signal set X=[x1,x2,…,x n ], where n is the number of data points, x i is the multi-source signal data at the i-th time point after preprocessing; Extract time domain and frequency domain features from the preprocessed signal; Step S3: Fusing features extracted from different signal sources to construct a multi-dimensional feature representation for characterizing the bearing state; Step S4: input the multi-dimensional features into a deep learning model that combines a convolutional neural network and a long short-term memory network for training. At the same time, the deep learning model introduces a multi-core maximum mean difference strategy to align the subspace distribution of input features under different working conditions; Convolutional neural networks are used to extract local spatial information of input features; Long short-term memory networks are used to model the dependencies between features over time; The extracted features are fed into a long short-term memory network consisting of multiple memory units through a fully connected layer for time series modeling. The Softmax function is then used at the output layer to convert the classification results into the predicted probabilities of various types of faults:

[0006] Where zk is the score of category k, K is the number of categories, is the predicted probability of class k, ; Among them, w i is the weighting coefficient, F i is the time-frequency characteristics of the i-th sensor; Step S5: Fault identification is performed on the real-time collected signals based on the trained deep learning model, and an alarm is issued when the identification result meets the preset conditions.

[0007] Furthermore, in step S2, time domain features are extracted including mean, standard deviation, skewness, and kurtosis features, and the time domain features are defined as:

[0008]

[0009]

[0010]

[0011] in, is the statistical mean of the signal segment, is the statistical standard deviation of the signal segment, is the skewness of signal X, is the kurtosis of signal X; The signal is analyzed in the frequency domain using fast Fourier transform to extract the spectrum mean, spectrum standard deviation, spectrum energy, and peak frequency characteristics.

[0012] Furthermore, in step S3, a domain adaptation strategy based on multi-kernel maximum mean difference is used. The domain adaptation strategy constructs a high-dimensional subspace mapping by combining multiple kernel functions of different types and scales.

[0013] Furthermore, the kernel function u is a positive definite function The linear combination of is represented by the weighted sum of multiple kernel functions:

[0014] in, is the i-th positive definite kernel function, is a positive definite kernel function The non-negative weight coefficients of .

[0015] Furthermore, in step S5, fault identification adopts a multi-label classification strategy, combined with the Binary Relevance strategy and the label correlation weight modeling mechanism, and incorporates the co-occurrence pattern between various fault labels into the joint modeling process. In the model deployment phase, a sliding time window mechanism is introduced to dynamically evaluate the continuously collected data. When the predicted probability of any category exceeds the set threshold within the window, or the concurrent probability of multiple categories meets the warning conditions at the same time, the system generates a composite fault alarm signal and performs priority classification prompts.

[0016] In a second aspect, the present application provides a press bearing fault diagnosis system, the system comprising: A data acquisition component is used to collect multi-source signal data of the press bearing during operation, wherein the multi-source signal data includes vibration signals, temperature signals and pressure signals; A preprocessing component, configured to perform denoising and standardization processing on the multi-source signals, generate a preprocessed time series signal set, and extract time domain and frequency domain features from the signal set; Feature fusion component, used to perform weighted fusion of time-frequency features from different sensors to construct a multi-dimensional feature representation for characterizing bearing status; A diagnostic model component for inputting the multi-dimensional features into a deep learning model combining a convolutional neural network and a long short-term memory network for training and inference. The deep learning model introduces a multi-core maximum mean difference strategy to adapt to feature migration issues under different operating conditions; The fault identification and alarm component is used to issue an alarm signal after the diagnostic model completes the classification and judgment of the real-time collected signals, when the predicted probability of any fault type meets the preset conditions, to prompt the operation and maintenance personnel to take maintenance measures.

[0017] In a third aspect, the present application provides a terminal, including: a memory for storing a diagnostic program for a press bearing failure; The processor is used to implement the steps of the press bearing fault diagnosis method as described in the first aspect when executing the press bearing fault diagnosis system.

[0018] In a fourth aspect, the present application provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the press bearing fault diagnosis method as described in the first aspect.

[0019] It can be seen from the above technical solutions that the advantages of the present invention are: (1) The deep fusion of multimodal data is achieved, which improves the accuracy and robustness of fault diagnosis. By introducing multimodal data fusion technology, data from multiple heterogeneous sources such as vibration sensors and temperature sensors are effectively integrated, avoiding the information loss and diagnostic bias that may be caused by a single data source. During the fusion process, by building a unified data expression framework and deep feature mapping mechanism, the correlation and complementarity between different modal data are enhanced, thereby achieving more comprehensive and accurate fault status identification, and significantly improving the adaptability and generalization ability of the diagnostic model under complex working conditions.

[0020] (2) The present invention is based on a deep learning framework, integrates time domain and frequency domain feature information, and uses structures such as convolutional neural networks and long short-term memory networks to extract key features from bearing fault signals. It can effectively identify a variety of typical fault types including inner ring faults, outer ring faults, rolling element faults and their combined faults. The diagnostic model has good generalization ability and can adapt to the variability characteristics under different loads, speeds and working conditions.

[0021] (3) This invention introduces an online monitoring and fault prediction mechanism, combined with sliding window processing and dynamic assessment methods, which can continuously sense changes in the health status of bearings, identify potential degradation trends in advance, and provide intelligent early warning and maintenance decision support for stamping equipment. This method effectively reduces unplanned downtime caused by sudden failures, extends equipment life, significantly reduces maintenance costs, and improves the overall operating efficiency and safety of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 The figure is a flow chart of a method for diagnosing a press bearing failure; Figure 2 Composite fault diagnosis framework diagram for press bearing fault diagnosis method. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] See also Figure 1 and Figure 2 As shown, the present invention provides a method for diagnosing a press bearing fault, comprising the following steps: Step S1, collecting multi-source signal data of a press bearing during operation, where the multi-source signal data includes a vibration signal, a temperature signal, and a pressure signal; Step S2: preprocessing the multi-source signals, including signal denoising and standardization operations; Get the preprocessed time series signal set X=[x1,x2,…,x n ], where n is the number of data points, x i is the multi-source signal data at the i-th time point after preprocessing; Extract time domain and frequency domain features from the preprocessed signal; Step S3: Fusing features extracted from different signal sources to construct a multi-dimensional feature representation for characterizing the bearing state; Step S4: input the multi-dimensional features into a deep learning model that combines a convolutional neural network and a long short-term memory network for training. At the same time, the deep learning model introduces a multi-core maximum mean difference strategy to align the subspace distribution of input features under different working conditions; Convolutional neural networks are used to extract local spatial information of input features; Long short-term memory networks are used to model the dependencies between features over time; The extracted features are fed into a long short-term memory network consisting of multiple memory units through a fully connected layer for time series modeling. The Softmax function is then used at the output layer to convert the classification results into the predicted probabilities of various types of faults:

[0026] Where zk is the score of category k, K is the number of categories, is the predicted probability of class k, ; Among them, w i is the weighting coefficient, F i is the time-frequency feature of the i-th sensor.

[0027] Step S5: Fault identification is performed on the real-time collected signals based on the trained deep learning model, and an alarm is issued when the identification result meets the preset conditions.

[0028] In some embodiments, in step S2, the time domain feature extraction includes mean, standard deviation, skewness, and kurtosis features, and the time domain features are defined as:

[0029]

[0030]

[0031]

[0032] in, is the statistical mean of the signal segment, is the statistical standard deviation of the signal segment, is the skewness of signal X, is the kurtosis of signal X; The signal is analyzed in the frequency domain using fast Fourier transform to extract the spectrum mean, spectrum standard deviation, spectrum energy, and peak frequency characteristics.

[0033] In some embodiments, in step S3, the signals collected by different sensors are , where m represents the number of sensors, and the time-frequency features extracted from each signal are F1, F2, ... F m , through the weighted fusion method, the time-frequency features of these signals are combined into a multidimensional feature vector: .

[0034] In some preferred embodiments, the deep learning model used in step S4 includes a combination of a convolutional neural network (CNN) and a long short-term memory network (LSTM) to achieve joint modeling of multi-dimensional features. Specifically: Convolutional neural networks are used to extract local spatial structural information from input features. The network structure consists of three convolutional layers connected in series, each with a 3×3 kernel size, which is used to extract local texture and variation patterns at different scales. Each convolutional layer is followed by a rectified linear unit (ReLU) activation function to enhance nonlinear representation capabilities. Max pooling is used to compress and downsample feature maps, reducing computational complexity and improving the hierarchical representation capabilities of features.

[0035] The feature map output by the convolutional layer undergoes dimension transformation and structural reconstruction through the fully connected layer to adapt to the input requirements of the subsequent sequence modeling network.

[0036] The long short-term memory network (LSTM) receives the feature vector sequence output by the fully connected layer and performs time series modeling on it to capture the dynamic correlation and evolution law of multidimensional features between different time steps, thereby enhancing the model's ability to identify the changing trends of the bearing's operating status.

[0037] To improve the model's generalization performance under complex operating conditions such as variable loads and speeds, the deep learning model further introduces a multi-kernel maximum mean discrepancy (MK-MMD) strategy to perform subspace mapping and distribution alignment of input features between the training domain and the target application domain. This MK-MMD mechanism minimizes cross-domain feature differences by constructing a linear combination of multiple positive definite kernel functions, thereby alleviating the decline in diagnostic accuracy caused by operating condition migration.

[0038] In order to adapt to the feature distribution deviation caused by changes in different working conditions (such as load and speed), the multi-kernel maximum mean difference (MK-MMD) strategy is introduced. By combining multiple kernel functions to construct a subspace mapping, the model's adaptability to inconsistent features between domains is enhanced.

[0039] The kernel function u is a positive definite function The linear combination of is represented by the weighted sum of multiple kernel functions:

[0040] in, is the i-th positive definite kernel function, is a positive definite kernel function The non-negative weight coefficients of .

[0041] At the model's output layer, a Softmax function is used to convert the neural network's output vector into predicted probabilities for each fault category, ensuring the classification results are both interpretable and discriminable. Specifically, each node in the model's output corresponds to a fault category, and its output value reflects the probability of occurrence of that category. This value can be used to determine whether a sample exhibits a particular or multiple fault characteristics.

[0042] Furthermore, in the final output layer of the model, the output of the model is converted into multi-category probabilities through the Softmax function:

[0043] Among them, z k is the score of category k, K is the number of categories, is the predicted probability of class k.

[0044] In some further embodiments of the present invention, the fault identification process described in step S5 employs a multi-label classification strategy to address the identification of multiple types of faults (such as inner race faults, outer race faults, and rolling element faults) that may exist simultaneously in press bearings during actual operation. This strategy supports the parallel determination of multiple potential fault modes, improving the system's ability to identify complex fault conditions.

[0045] Specifically, this embodiment represents the fault status of each sample as a label vector, where each dimension corresponds to a predefined fault type. The system employs a Binary Relevance (BR) strategy, breaking the multi-label classification task into several independent binary classification subtasks and training a corresponding sub-classifier for each fault type.

[0046] Construct the output layer structure:

[0047] Where C is the number of fault categories, and the model can identify multiple fault labels at the same time.

[0048] During the training process, the cross entropy loss function is used as the objective function, the optimization method is the Adam optimizer, and a validation set is set for overfitting control and parameter tuning. The training goal is to minimize:

[0049] where p i is the predicted probability of class i.

[0050] On this basis, in order to improve the model's ability to express the dependency relationship between labels, a label correlation modeling mechanism is introduced. By introducing parameters such as label co-occurrence probability and mutual information weight, the coupling characteristics between multiple fault types are captured, thereby improving the accuracy of compound fault identification.

[0051] To enhance the stability and robustness of the diagnostic system in engineering applications, this embodiment introduces a sliding time window mechanism during the model deployment phase. Specifically, the real-time sensor data is continuously segmented into fixed window lengths (e.g., 5 seconds, 1024 frames), and the prediction results in each window are dynamically accumulated and evaluated. The system sets the following fault alarm triggering rules: When the cumulative predicted probability of a certain fault type within the sliding time window continuously exceeds the set threshold (e.g., 0.85), the fault type is considered to have a high confidence level. The system will determine that the fault type exists and issue an alarm. Alternatively, within the same time window, if the predicted probabilities of multiple fault categories simultaneously exceed the corresponding thresholds, the system will determine it as a composite fault state and prioritize the faults based on the predicted probabilities and weight calculation results of each category to assist operation and maintenance personnel in arranging the maintenance sequence according to the severity.

[0052] In some embodiments, the present application provides a system for press bearing fault diagnosis, which is suitable for real-time monitoring and intelligent diagnosis of bearing operating conditions in high-speed stamping production lines. The system includes: A data acquisition component is used to synchronously collect multi-source signal data related to the bearing during the operation of the press. The multi-source signals include but are not limited to vibration signals, temperature signals, and pressure signals. The vibration signal is used to reflect the dynamic response characteristics of the machine, and the temperature and pressure signals are used to provide auxiliary information on the operating environment and load status. The three types of signals are synchronously acquired at a preset sampling frequency via different types of industrial sensors; The preprocessing component is used to standardize the collected multi-source signals. The preprocessing process includes noise suppression, outlier removal, and amplitude normalization. After preprocessing, the system divides each type of signal into time windows of equal length to construct a unified time series signal set. This component further extracts typical statistical and spectral features from each signal segment, including mean, standard deviation, skewness, kurtosis, spectral energy, and peak frequency, to characterize the state changes of the bearing over different time periods. A feature fusion component is used to perform weighted fusion of features from multiple sensors to form a multi-dimensional feature vector representing the current bearing state. The weighting coefficients can be set based on sensor reliability, historical performance, or learning results from the training phase to enhance the discriminative power and robustness of the fused features. The diagnostic model component is used to perform classification learning and fault identification on the fused feature vectors. This model uses a composite structure combining a convolutional neural network (CNN) and a long short-term memory network (LSTM). The CNN is used to extract local spatial features of the input features, and the LSTM is used to model the dependencies of features evolving over time. To adapt to distribution shifts caused by changes in various operating conditions, the model further introduces a multi-kernel maximum mean difference (MK-MMD) strategy to align the feature distributions of the source and target domains, thereby improving generalization capabilities under unseen operating conditions. The fault identification and alarm component is used to deploy the trained deep learning model to the on-site diagnostic platform, receive real-time collected signals and perform fault identification operations; this component supports multi-label classification strategies, and each fault mode has an independent output node. When the predicted probability corresponding to any fault label continues to exceed the set threshold (for example, 0.85) within the sliding time window, the system will automatically issue an alarm signal and send maintenance prompts to operators through a graphical interface or industrial communication protocol to assist them in making predictions and handling faults.

[0053] In some embodiments, the present application provides a terminal, including: a memory for storing a diagnostic program for a press bearing failure; A processor is used to implement the steps of the press bearing fault diagnosis method when executing the press bearing fault diagnosis system.

[0054] In some embodiments, the present application provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the press bearing fault diagnosis method.

[0055] It is understood that the systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or physical devices, or by products having certain functions. A typical implementation device is a computer, which may be a personal computer, a laptop computer, a personal digital assistant, a tablet computer, a wearable device, or a combination of any of these devices.

[0056] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0057] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0058] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be used to store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0059] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0060] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0061] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0062] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms "a," "an," "the," and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0063] It should be understood that although the terms first, second, third, etc. may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when..." or "when..." or "in response to determining."

[0064] The above description is merely a preferred embodiment of one or more embodiments of this specification and is not intended to limit one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of this specification shall be included in the scope of protection of one or more embodiments of this specification.

Claims

1. A method for diagnosing a press bearing fault, characterized in that: The following steps are involved: Step S1, collecting multi-source signal data of a press bearing during operation, where the multi-source signal data includes a vibration signal, a temperature signal, and a pressure signal; Step S2: preprocessing the multi-source signals, including signal denoising and standardization operations; Get the preprocessed time series signal set X=[x1,x2,…,x n ], where n is the number of data points, x i is the multi-source signal data at the i-th time point after preprocessing; Extract time domain and frequency domain features from the preprocessed signal; Step S3: Fusing features extracted from different signal sources to construct a multi-dimensional feature representation for characterizing the bearing state; Step S4: input the multi-dimensional features into a deep learning model that combines a convolutional neural network and a long short-term memory network for training. At the same time, the deep learning model introduces a multi-core maximum mean difference strategy to align the subspace distribution of input features under different working conditions; Convolutional neural networks are used to extract local spatial information of input features; Long short-term memory networks are used to model the dependencies between features over time; The extracted features are fed into a long short-term memory network consisting of multiple memory units through a fully connected layer for time series modeling. The Softmax function is then used at the output layer to convert the classification results into the predicted probabilities of various types of faults: Where zk is the score of category k, K is the number of categories, is the predicted probability of class k, ; Among them, w i is the weighting coefficient, F i is the time-frequency characteristics of the i-th sensor; Step S5: Fault identification is performed on the real-time collected signals based on the trained deep learning model, and an alarm is issued when the identification result meets the preset conditions.

2. The method for diagnosing a press bearing fault according to claim 1, characterized in that: In step S2, the time domain feature extraction includes mean, standard deviation, skewness, and kurtosis features, and the time domain features are defined as: in, is the statistical mean of the signal segment, is the statistical standard deviation of the signal segment, is the skewness of signal X, is the kurtosis of signal X; The signal is analyzed in the frequency domain using fast Fourier transform to extract the spectrum mean, spectrum standard deviation, spectrum energy, and peak frequency characteristics.

3. The method for diagnosing a press bearing fault according to claim 1, characterized in that: In step S3, a domain adaptation strategy based on multi-kernel maximum mean difference is used. The domain adaptation strategy constructs a high-dimensional subspace mapping by combining multiple kernel functions of different types and scales.

4. The method for diagnosing a press bearing fault according to claim 3, characterized in that: The kernel function u is a positive definite function The linear combination of is represented by the weighted sum of multiple kernel functions: in, is the i-th positive definite kernel function, is a positive definite kernel function The non-negative weight coefficients of .

5. The method for diagnosing a press bearing fault according to claim 1, characterized in that: In step S5, fault identification adopts a multi-label classification strategy, combined with the Binary Relevance strategy and the label correlation weight modeling mechanism, and incorporates the co-occurrence pattern between various fault labels into the joint modeling process. In the model deployment phase, a sliding time window mechanism is introduced to dynamically evaluate continuously collected data. When the predicted probability of any category exceeds the set threshold within the window, or the concurrent probability of multiple categories meets the warning conditions at the same time, the system generates a composite fault alarm signal and provides a priority classification prompt.

6. A diagnostic system for press bearing failure, characterized in that: The system includes: A data acquisition component is used to collect multi-source signal data of the press bearing during operation, wherein the multi-source signal data includes vibration signals, temperature signals and pressure signals; A preprocessing component, configured to perform denoising and standardization processing on the multi-source signals, generate a preprocessed time series signal set, and extract time domain and frequency domain features from the signal set; Feature fusion component, used to perform weighted fusion of time-frequency features from different sensors to construct a multi-dimensional feature representation for characterizing bearing status; A diagnostic model component for inputting the multi-dimensional features into a deep learning model combining a convolutional neural network and a long short-term memory network for training and inference. The deep learning model introduces a multi-core maximum mean difference strategy to adapt to feature migration issues under different operating conditions; The fault identification and alarm component is used to issue an alarm signal after the diagnostic model completes the classification and judgment of the real-time collected signals, when the predicted probability of any fault type meets the preset conditions, to prompt the operation and maintenance personnel to take maintenance measures.

7. A terminal, characterized in that: include: a memory for storing a diagnostic program for a press bearing failure; A processor is used to implement the steps of the press bearing fault diagnosis method as claimed in claim 1 when executing the press bearing fault diagnosis system.

8. A computer-readable storage medium, characterized in that The storage medium stores computer instructions. When the computer reads the computer instructions in the storage medium, the computer executes the press bearing fault diagnosis method as claimed in claim 1.

Images