Linear Guide Fault Diagnosis Method and System Based on Deep Learning

Abstract

The present invention relates to the technical field of operation and maintenance management of linear guides, and particularly to a method and system for fault diagnosis of linear guides based on deep learning, which can improve the stability and production efficiency of equipment and reduce maintenance costs. The method includes: obtaining first operation state data of a target linear guide and second operation state data of at least one associated linear guide; inputting the first operation state data into a preset guide fault evaluation model to obtain a first fault evaluation result; inputting at least one set of second operation state data into the guide fault evaluation model respectively to obtain at least one set of second fault evaluation results; considering the influence of homologous faults, correcting the first fault evaluation result according to at least one set of second fault evaluation results to obtain a target linear guide fault diagnosis result.

Description

Technical Field

[0001] The present invention relates to the technical field of operation and maintenance management of linear guides, and particularly to a method and system for fault diagnosis of linear guides based on deep learning. Background Art

[0002] In a linear guide system, fault diagnosis is a key link to ensure the stable operation of mechanical equipment and maintain production efficiency; as a precision mechanical component, linear guides are widely used in various automation equipment, such as numerically controlled machine tools, industrial robots, etc.; due to long-term load-bearing, friction and vibration, linear guides are prone to various faults, which not only affect the accuracy and stability of the equipment, but may also cause equipment downtime and production losses.

[0003] Most of the existing linear guide fault diagnosis methods are for individual equipment or components, ignoring the mutual correlation and fault propagation characteristics between equipment in the actual industrial scenario. Especially in a linear guide system, adjacent linear guides are often in similar operating environments and conditions, and a fault in one guide may have a homologous fault impact on its adjacent guides. This homologous fault impact increases the complexity of fault diagnosis, making it difficult for traditional methods to accurately judge the true source and degree of the fault. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a method and system for fault diagnosis of linear guides based on deep learning, which can improve the stability and production efficiency of equipment and reduce maintenance costs.

[0005] In a first aspect, the present invention provides a method for fault diagnosis of linear guides based on deep learning, the method comprising:

[0006] Obtaining first operating state data of a target linear guide and second operating state data of at least one associated linear guide;

[0007] Inputting the first operating state data into a preset guide fault evaluation model to obtain a first fault evaluation result;

[0008] Inputting at least one set of the second operating state data into the guide fault evaluation model respectively to obtain at least one set of second fault evaluation results;

[0009] Considering the homologous fault impact, correcting the first fault evaluation result according to at least one set of the second fault evaluation results to obtain a target linear guide fault diagnosis result.

[0010] Further, the associated linear guide is a linear guide of the same type as the target linear guide and adjacent in position.

[0011] Further, a method for collecting operating state data includes:

[0012] Install a variety of sensors on the linear guide and adjacent linear guides for real-time monitoring of their operating states; the sensors include a temperature sensor, an acceleration sensor, a vibration sensor, and a noise sensor;

[0013] Set the data collection frequency;

[0014] According to the data collection frequency, collect data in real time through the sensors installed on the linear guide;

[0015] Preprocess the collected sensor data;

[0016] Store the preprocessed sensor data in a preset database.

[0017] Further, the first fault assessment result includes a ball wear fault probability value, a guide corrosion fault probability value, a slider deformation fault probability value, and a slider end cover detachment fault probability value.

[0018] Further, a method for constructing the guide fault assessment model includes:

[0019] Collect the operating state data of historical linear guides and preprocess the collected data;

[0020] Select a deep learning model as the basic architecture of the guide fault assessment model; the deep learning model includes a recurrent neural network and a convolutional neural network;

[0021] Divide the preprocessed operating state data of historical linear guides into a training set, a validation set, and a test set;

[0022] Use the training set to train the model and adjust the parameters of the model;

[0023] Use the validation set data to validate the model and adjust the parameters and structure of the model according to the validation results;

[0024] Use the test set data to evaluate the model;

[0025] Deploy the trained model to the production environment, receive the real-time data stream from the sensors, and perform online fault assessment.

[0026] Further, the mathematical calculation formula for correcting the first fault assessment result according to at least one set of the second fault assessment results is: Where P c represents the corrected fault probability value; P trepresents the original target guide rail fault probability value; α is an adjustment coefficient between 0 and 1, used to balance the proportion of the original assessment and the influence of homologous faults; n represents the number of associated guide rails; w i represents the weight of the i-th associated guide rail; P ai represents the fault probability value of the i-th associated guide rail.

[0027] Furthermore, the target linear guide rail fault diagnosis result includes the corrected fault probability value, fault type confirmation, fault severity assessment, homologous fault influence analysis, and maintenance suggestions.

[0028] On the other hand, the present application also provides a linear guide rail fault diagnosis system based on deep learning, and the system includes:

[0029] A data acquisition module that acquires the first operating state data of the target linear guide rail and the second operating state data of at least one associated linear guide rail;

[0030] A guide rail fault assessment module that inputs the first operating state data into a preset guide rail fault assessment model to obtain a first fault assessment result;

[0031] An associated guide rail fault assessment module that inputs at least one set of the second operating state data into the guide rail fault assessment model respectively to obtain at least one set of second fault assessment results;

[0032] A fault diagnosis module that corrects the first fault assessment result according to at least one set of the second fault assessment results considering the influence of homologous faults to obtain the target linear guide rail fault diagnosis result.

[0033] In a third aspect, the present application provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are connected through the bus, and when the computer program is executed by the processor, the steps in any one of the above methods are implemented.

[0034] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps in any one of the above methods are implemented.

[0035] The beneficial effects of the present invention compared with the prior art are as follows: This method not only evaluates the state of a single linear guide rail, but also analyzes the mutual influence between adjacent guide rails; by obtaining and processing the operating state data of at least one associated linear guide rail, the influence of homologous faults can be more accurately identified, thereby improving the diagnostic accuracy; by comprehensively analyzing the data of the target guide rail and the associated guide rails and considering the influence of homologous faults to correct the fault assessment results, the location, type and severity of the faults can be more precisely judged, reducing the possibility of misdiagnosis or missed diagnosis;

[0036] Accurate fault diagnosis helps to plan maintenance activities in advance, making the maintenance work more predictable and targeted, reducing unexpected downtime, lowering maintenance costs, and improving production efficiency;

[0037] The deep learning model can handle complex non-linear relationships and is suitable for the recognition of different types of fault modes; even in complex working conditions such as long-term load, friction and vibration, it can maintain good performance; with the accumulation of more operating state data, the preset guide rail fault assessment model can self-optimize through continuous learning and training, improving the accuracy and reliability of its diagnosis over time;

[0038] The entire diagnostic process from data acquisition to the generation of the final fault diagnosis result is automated, greatly reducing the need for manual intervention, improving the system's response speed and working efficiency, and at the same time reducing the risk of human errors;

[0039] The detailed fault diagnosis results provided by the method can be directly used to guide maintenance personnel to make correct maintenance decisions, ensuring the stable operation and high-efficiency production of mechanical equipment; through accurate fault diagnosis, enterprises can avoid unnecessary comprehensive inspections and over-maintenance, saving human and material resources, and at the same time improving the availability and service life of the equipment;

[0040] In summary, the above-mentioned linear guide rail fault diagnosis method based on deep learning overcomes the deficiencies of traditional fault diagnosis methods, can improve the stability and production efficiency of equipment, and reduce maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a flowchart of the present invention;

[0042] Figure 2 is a flowchart of the construction method of the guide rail fault assessment model;

[0043] Figure 3 is a structural diagram of a linear guide rail fault diagnosis system based on deep learning;

[0044] Figure 4 is a sample configuration diagram under four different working conditions. Detailed Implementation Modes

[0045] In the description of the present application, those skilled in the art should know that the present application can be implemented as a method, a device, an electronic device, and a computer-readable storage medium. Therefore, the present application can be specifically implemented in the following forms: complete hardware, complete software (including firmware, resident software, microcode, etc.), and a combination of hardware and software. In addition, in some embodiments, the present application can also be implemented in the form of a computer program product in one or more computer-readable storage media, and the computer-readable storage medium contains computer program code.

[0046] The above-mentioned computer-readable storage media can adopt any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination of the above. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, flash memories, optical fibers, compact discs read-only memories, optical storage devices, magnetic storage devices, or any combination of the above. In the present application, the computer-readable storage media can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or component.

[0047] In the technical solution of the present application, the acquisition, storage, use, processing, etc. of data all comply with the relevant provisions of national laws.

[0048] The present application describes the provided method, device, and electronic device through flowcharts and / or block diagrams.

[0049] It should be understood that each block of the flowchart and / or block diagram, as well as the combination of blocks in the flowchart and / or block diagram, can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing devices, thereby producing a machine. These computer-readable program instructions are executed by a computer or other programmable data processing devices, resulting in a device that realizes the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0050] These computer-readable program instructions can also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to work in a specific manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction device product that includes the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0051] Computer-readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other devices, causing a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other devices to generate a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus can provide a process for implementing the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0052] The present application will be described below with reference to the accompanying drawings in the present application.

[0053] Embodiment 1: As Figures 1 to 2 shown, the linear guide fault diagnosis method based on deep learning of the present invention specifically includes the following steps:

[0054] S1. Obtain the first operating state data of the target linear guide and the second operating state data of at least one associated linear guide;

[0055] The associated linear guide is a linear guide of the same type as the target linear guide and adjacent in position; the data content of the first operating state data and the second operating state data is the same, and both include at least any one of temperature data, acceleration data, vibration data, and noise data;

[0056] The method for collecting operating state data includes:

[0057] Install a variety of sensors on the linear guide and the adjacent linear guide for real-time monitoring of its operating state; the sensors include a temperature sensor, an acceleration sensor, a vibration sensor, and a noise sensor;

[0058] The temperature sensor is used to monitor the temperature change of the linear guide and the surrounding environment. An abnormal temperature rise is a sign of wear and insufficient lubrication;

[0059] The acceleration sensor can detect sudden movements or impacts caused by friction or mechanical problems;

[0060] The vibration sensor is used to record the vibration mode generated during the operation of the device, which helps to identify mechanical defects or imbalances;

[0061] The noise sensor can capture abnormal sounds caused by faults, such as abnormal friction sounds or impact sounds between metals;

[0062] Set the data acquisition frequency to monitor the change in the long-term operating state of the linear guide;

[0063] According to the data acquisition frequency, collect data in real time through the sensors installed on the linear guide to ensure the timeliness and accuracy of the data;

[0064] Preprocess the collected sensor data to ensure the quality and usability of the data; this includes data cleaning, data format conversion, and data normalization steps;

[0065] Store the preprocessed sensor data in a preset database for subsequent analysis and processing; to ensure the security and reliability of the data, take data backup and recovery measures for the stored data through the database.

[0066] In this step, by obtaining the first operating state data and the second operating state data of the target linear guide and at least one associated linear guide, comprehensive monitoring of the linear guide system is achieved, which can reflect the operating state of the linear guide in all directions, helping to detect potential problems in a timely manner; due to the introduction of the associated linear guide, this step can more accurately locate the problem; when an abnormal situation occurs in the target linear guide, by comparing its data with that of the associated linear guide, it is possible to more quickly determine whether it is a local problem or a systematic problem, and thus take targeted maintenance measures; by installing a variety of sensors on the linear guide and the adjacent linear guide and setting the data acquisition frequency, real-time monitoring of the operating state of the linear guide is achieved; it helps to detect abnormal situations in a timely manner and issue early warnings to avoid the further expansion of faults and ensure the stable operation of the equipment; preprocessing the collected sensor data ensures the quality and usability of the data; at the same time, storing the preprocessed data in a preset database provides reliable data support for subsequent analysis and processing; through data backup and recovery measures, the security and reliability of the data are further guaranteed; through the implementation of this step, potential problems of the linear guide can be discovered and processed in a timely manner, thus avoiding the extension of downtime and the increase in maintenance costs caused by faults; at the same time, through real-time monitoring and data analysis, the maintenance plan can also be optimized, the maintenance efficiency can be improved, and the overall maintenance cost can be reduced.

[0067] S2. Input the first operating state data into a preset guide rail fault assessment model to obtain a first fault assessment result;

[0068] The first fault assessment result includes the probability value of ball wear fault, the probability value of guide rail corrosion fault, the probability value of slider deformation fault, and the probability value of slider end cover detachment fault;

[0069] The probability value of ball wear fault is used to evaluate the possibility of excessive wear of the balls;

[0070] The probability value of guide rail corrosion fault is used to judge the probability of the existence of corrosion on the surface of the guide rail;

[0071] The probability value of slider deformation fault is used to detect whether there is abnormal deformation of the slider;

[0072] The probability value of the slider end cover falling off is used to predict the risk of loosening or falling off of the slider end cover;

[0073] The method for constructing the guide rail fault evaluation model includes:

[0074] Collect the operating status data of historical linear guide rails, including temperature data, acceleration data, vibration data, and noise data;

[0075] Preprocess the collected data, including data cleaning, data format conversion, and data normalization steps to ensure the quality and consistency of the data;

[0076] Extract features from the preprocessed operating status data of historical linear guide rails to extract features that can reflect the faults of linear guide rails;

[0077] Select a deep learning model as the basic architecture of the guide rail fault evaluation model; the deep learning model includes a recurrent neural network and a convolutional neural network;

[0078] Design the input layer, hidden layer, and output layer of the model; the input layer is used to receive the preprocessed data, the hidden layer is used for feature extraction and pattern recognition, and the output layer is used to output the fault evaluation result; in the hidden layer, design multiple convolutional layers, pooling layers, fully connected layers, etc. to extract deeper features;

[0079] Divide the preprocessed operating status data of historical linear guide rails into a training set, a validation set, and a test set; the training set is used for model training, the validation set is used for model validation and adjustment, and the test set is used to evaluate the performance of the model;

[0080] Use the training set to train the model, adjust the parameters of the model through the backpropagation algorithm to minimize the error between the prediction result of the model and the actual fault type; during the training process, use an optimization algorithm to accelerate the training process and improve the convergence of the model; optimize the performance of the model by adjusting the hyperparameters of the model;

[0081] Use the validation set data to validate the model, and adjust the parameters and structure of the model according to the validation results; repeat the training and optimization process until the performance of the model on the validation set reaches the optimal;

[0082] Use the test set data to evaluate the model, calculate indicators such as the accuracy rate, recall rate, and F1 score of the model; by comparing the evaluation results of different models, select the model with the best performance as the final guide rail fault evaluation model;

[0083] Deploy the trained model to the production environment, receive the real-time data stream from the sensor, and perform online fault evaluation.

[0084] In this step, by constructing a guide rail fault assessment model based on deep learning, the operation status data of historical linear guide rails can be fully utilized, and through the comprehensive processing and analysis of information, the model can more accurately predict fault types such as ball wear, guide rail corrosion, slider deformation, and slider end cover detachment, improving the accuracy of fault prediction; the fault probability values output by the model provide clear fault warning information for maintenance personnel; it helps maintenance personnel take measures in advance to prevent the occurrence of faults, or detect and handle them in a timely manner at the initial stage of the fault, thereby avoiding equipment downtime and production losses caused by the expansion of the fault; the online fault assessment based on the model enables maintenance personnel to understand the operation status of the linear guide rail in real time, reducing the frequency and intensity of manual inspections; at the same time, the model can automatically identify the fault type, reducing the time for maintenance personnel to judge the cause of the fault and improving the maintenance efficiency; in addition, by preventing the occurrence of faults, the maintenance cost and production interruption cost caused by faults are reduced; through the introduction of the deep learning model, the intelligent monitoring and fault prediction of the operation status of the linear guide rail are realized, providing strong support for building an intelligent operation and maintenance management system; in this step, by constructing a guide rail fault assessment model based on deep learning, the intelligent monitoring and fault prediction of the operation status of the linear guide rail are realized, improving the accuracy, maintenance efficiency, and intelligent level of fault prediction, and reducing the operation and maintenance cost and production interruption risk.

[0085] S3. Input at least one group of the second operation status data into the guide rail fault assessment model respectively to obtain at least one group of second fault assessment results; the second fault assessment results include the ball wear fault probability value, the guide rail corrosion fault probability value, the slider deformation fault probability value, and the slider end cover detachment fault probability value;

[0086] Preprocess the second operation status data, including data cleaning, data format conversion, and data normalization;

[0087] Input the preprocessed second operation status data into a preset guide rail fault assessment model; the model can automatically extract the features in the data and perform probability prediction of the fault type;

[0088] The model outputs at least one group of second fault assessment results, and each group of results corresponds to an associated linear guide rail; the results include the ball wear fault probability value, the guide rail corrosion fault probability value, the slider deformation fault probability value, and the slider end cover detachment fault probability value; the probability value reflects the possibility of the associated linear guide rail having the corresponding fault type;

[0089] Record and store the output fault assessment results in a database for subsequent analysis and comparison.

[0090] In this step, by collecting and preprocessing the second operating state data of at least one set of associated linear guides, the operating conditions of these guides can be comprehensively obtained; through preprocessing steps such as data cleaning, format conversion, and normalization, the quality and consistency of the data are ensured, providing a reliable basis for subsequent analysis; the preprocessed data is input into a preset guide fault assessment model, which is constructed based on deep learning and can automatically extract key features from complex data and perform probability prediction of fault types, greatly improving the efficiency and accuracy of fault diagnosis; the second fault assessment result output by the model intuitively reflects the possibility of the corresponding fault type occurring in the associated linear guides, providing clear fault warning information for maintenance personnel; maintenance personnel can compare the fault assessment results of the target linear guide and the associated linear guides, analyze the propagation and impact of homologous faults, and thus more accurately judge the true source and degree of the fault; based on detailed fault probability assessment and analysis of the impact of homologous faults, maintenance personnel can formulate more scientific and reasonable maintenance plans and fault handling strategies; this helps to reduce equipment downtime caused by faults, improve production efficiency and equipment utilization rate, and reduce maintenance costs.

[0091] S4. Considering the influence of homologous faults, correct the first fault assessment result according to at least one set of the second fault assessment results to obtain the fault diagnosis result of the target linear guide;

[0092] Analyze the second fault assessment results of the associated linear guides. By comparing the distribution of fault probability values and the proportion of fault types of each associated guide, identify whether there are similar fault patterns; if multiple associated guides show high probability values for the same or similar fault types, it indicates the influence of homologous faults;

[0093] According to the identified homologous fault pattern, correct the first fault assessment result of the target linear guide;

[0094] If it is determined that there are homologous faults and the fault probability value of the associated guide is higher than the corresponding fault probability value of the target guide, then consider appropriately increasing the fault probability value of the target guide to reflect the additional risk brought by the homologous faults;

[0095] Conversely, if the fault probability value of the associated guide is low and this low probability value is due to homologous factors, then carefully evaluate the true fault situation of the target guide;

[0096] During the correction process, different weights are assigned to the fault assessment results of different associated guides; the assignment of weights is based on factors such as the physical distance between the associated guide and the target guide, the similarity of the operating environment, and the difference in working conditions;

[0097] By using the method of weighted average, combine the fault probability values of each associated guide and the corresponding weights to correct the fault assessment result of the target guide;

[0098] The mathematical calculation formula for correcting the first fault evaluation result according to at least one set of the second fault evaluation results is as follows: where P c represents the corrected fault probability value; P t represents the original target guide rail fault probability value; α is an adjustment coefficient between 0 and 1, which is used to balance the proportion of the original evaluation and the influence of homologous faults; n represents the number of associated guide rails; w i represents the weight of the i-th associated guide rail; P ai represents the fault probability value of the i-th associated guide rail;

[0099] The target linear guide rail fault diagnosis results include:

[0100] The corrected fault probability value: For each fault type, the corrected fault probability value gives the possibility of the target linear guide rail having a fault in this fault type; this probability value has considered the influence of homologous faults, so it is more accurate than the original first fault evaluation result;

[0101] Fault type confirmation: According to the corrected fault probability value, the fault type that the target linear guide rail is about to have can be determined; it helps the operation and maintenance management personnel quickly locate the fault point and take corresponding maintenance measures;

[0102] Fault severity assessment: Combining historical data and experience, the severity of the fault can be initially evaluated; it helps the operation and maintenance management personnel understand the specific impact of the fault on the equipment operation and production efficiency;

[0103] Analysis of the influence of homologous faults: During the correction process, by comparing the fault probability values of the target linear guide rail and the associated linear guide rails, the influence degree of homologous faults on the target guide rail can be analyzed; it helps the operation and maintenance management personnel understand the propagation characteristics of the faults and the mutual influence between adjacent devices, and provides a basis for formulating preventive measures;

[0104] Maintenance suggestions: Based on the fault diagnosis results, the operation and maintenance management personnel can formulate targeted maintenance plans and suggestions.

[0105] In this step, by comprehensively correlating the fault evaluation results of the linear guide, the influence of homologous faults on the fault probability value of the target guide can be identified and corrected, thus more accurately reflecting the true fault situation of the target guide; it helps to improve the accuracy and reliability of fault diagnosis; after considering the influence of homologous faults, potential faults can be predicted and identified earlier, providing sufficient time for operation and maintenance management personnel to carry out maintenance and repair work, avoiding the expansion of faults or causing equipment downtime; accurate fault diagnosis results provide a scientific decision-making basis for operation and maintenance management personnel, helping to optimize operation and maintenance strategies and resource allocation, reducing operation and maintenance costs, and improving production efficiency; by timely correcting and predicting faults, this step helps to enhance the stability and reliability of the entire linear guide system, reducing production interruptions and losses caused by faults; the correction formula and weight allocation method in this step have a certain degree of flexibility and scalability, and can be adjusted and optimized according to the actual industrial scenario and the characteristics of the linear guide system to adapt to different fault diagnosis requirements; this step can significantly improve the accuracy, prediction ability and operation and maintenance management efficiency of fault diagnosis, providing a strong guarantee for the stable operation and production efficiency of the linear guide system.

[0106] Embodiment 2: As Figure 3 shown, the linear guide fault diagnosis system based on deep learning of the present invention specifically includes the following modules;

[0107] A data acquisition module that acquires the first operating state data of the target linear guide and the second operating state data of at least one associated linear guide;

[0108] A guide rail fault evaluation module that inputs the first operating state data into a preset guide rail fault evaluation model to obtain a first fault evaluation result;

[0109] An associated guide rail fault evaluation module that inputs at least one set of the second operating state data into the guide rail fault evaluation model respectively to obtain at least one set of second fault evaluation results;

[0110] A fault diagnosis module that, considering the influence of homologous faults, corrects the first fault evaluation result according to at least one set of the second fault evaluation results to obtain a fault diagnosis result of the target linear guide.

[0111] This system not only focuses on the operating state of the target linear guide, but also simultaneously collects and analyzes the data of the linear guide associated with the target linear guide, which helps to capture the influence of homologous faults and improve the accuracy of fault diagnosis;

[0112] By introducing the impact analysis of homologous faults, the fault diagnosis module can more accurately determine the source and degree of faults, avoiding misdiagnosis or missed diagnosis that may be caused by ignoring the correlation between devices; accurate fault diagnosis results can help maintenance personnel plan maintenance activities in advance, reduce unexpected downtime, lower production losses, and optimize the allocation of maintenance resources;

[0113] Deep learning models can handle non-linear and complex input-output relationships and are suitable for different types of fault mode recognition; as more operating state data is collected and analyzed, the guide rail fault assessment model can self-optimize through continuous learning and training, thereby improving the accuracy and reliability of fault diagnosis over time;

[0114] The entire process from data acquisition to the generation of fault diagnosis results is automated, reducing the need for manual intervention, improving the system's response speed and working efficiency, and simultaneously reducing the risk of human errors; it provides intuitive and detailed fault diagnosis information for maintenance management personnel, helping them make more informed maintenance decisions and ensuring the stable operation and efficient production of mechanical equipment;

[0115] By integrating deep learning technology and the understanding of the interaction between linear guide rails, the system provides a more intelligent, accurate, and efficient fault diagnosis solution, which helps improve the stability and production efficiency of equipment and reduce maintenance costs.

[0116] Embodiment 3: As Figure 4 shown, taking the acquisition of acceleration data of the target linear guide rail as an example, the steps for fault diagnosis of the target linear guide rail include:

[0117] Set the sampling frequency of the acceleration data, and collect three groups of data for eight seconds respectively under different rotational speeds and loads; in order to simulate the real operating environment of the linear guide rail and the state when a fault occurs, in this embodiment, each fault type was tested under four different working conditions;

[0118] Determine the sampling points of each sensor, and put the data collected by the sensors into two channels; the vibration data collected by the acceleration sensor often contains a large amount of noise, and directly using it will reduce the fault recognition rate of the model, so it is necessary to preprocess the collected data; decomposing the signal into different frequency band modes can effectively separate the noise and useful information in the signal; when decomposing the vibration data, it is necessary to determine the number of modes and the penalty parameter. If the number of modes is set too large, it will lead to over-decomposition and mode mixing; setting it too small will cause under-decomposition and loss of characteristic information;

[0119] Divide the training set, validation set, and test set according to the fault category, and normalize the converted data; divide the converted data according to the fault category, with a division ratio of 8:1:1; the training set is used for iterative training of the model, the test set is used to evaluate the change in the model accuracy during the training process, and the validation set is used to evaluate the generalization effect of the model;

[0120] Input the data in the training set into the model for training, and determine the model parameters by the validation set and the test set; the model includes a multi-channel extraction module, a feature screening module, and a fault classification module;

[0121] Optimize the number of modes and the penalty parameter, and process the original data. Since the obtained mode functions contain different amounts of information and noise, they need to be selected; when the mode function contains more fault feature information, the sparsity of the signal is stronger and the envelope entropy is smaller, and vice versa, the envelope entropy is larger. Therefore, the minimum envelope entropy is used as the standard for screening the mode function, and the three selected mode components are input into the model for training; the convolutional layer extracts the feature information, the pooling layer reduces the size of the feature map to reduce the number of calculation parameters, and the residual connection can alleviate the problems of gradient disappearance and explosion during gradient backpropagation, and can also speed up the training speed and improve the model performance; finally, the feature information is flattened and input through the fully connected layer, and the fault classification label is output;

[0122] Use the trained model to conduct fault diagnosis experiments under different working conditions.

[0123] All the various change methods and specific embodiments of the linear guide fault diagnosis method based on deep learning in the foregoing Embodiment 1 are equally applicable to the linear guide fault diagnosis system based on deep learning in this embodiment. Through the foregoing detailed description of the linear guide fault diagnosis method based on deep learning, those skilled in the art can clearly know the implementation method of the linear guide fault diagnosis system based on deep learning in this embodiment. Therefore, for the sake of simplicity of the specification, it will not be elaborated here.

[0124] In addition, the present application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected through the bus. When the computer program is executed by the processor, it implements each process of the method embodiment for controlling the output data, and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0125] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A linear guide fault diagnosis method based on deep learning, characterized in that, The method includes: Obtaining the first operating state data of the target linear guide and the second operating state data of at least one associated linear guide; Inputting the first operating state data into a preset guide rail fault assessment model to obtain a first fault assessment result; Inputting at least one set of the second operating state data into the guide rail fault assessment model respectively to obtain at least one set of second fault assessment results; Considering the influence of homologous faults, correcting the first fault assessment result according to at least one set of the second fault assessment results to obtain a fault diagnosis result of the target linear guide; The associated linear guide is a linear guide of the same type as the target linear guide and adjacent in position; During the correction process, different weights are assigned to the second fault assessment results of different associated linear guides, and the weight assignment is determined based on the physical distance between the associated linear guide and the target linear guide, the similarity of the operating environment, and the difference degree of the working conditions.

2. The method for diagnosing linear guide rail faults based on deep learning according to claim 1, characterized in that, The method for collecting operating state data includes: Installing various sensors on the linear guide and the adjacent linear guide for real-time monitoring of their operating states; the sensors include temperature sensors, acceleration sensors, vibration sensors, and noise sensors; Setting the data collection frequency; According to the data collection frequency, collecting data in real time through the sensors installed on the linear guide; Preprocessing the collected sensor data; Storing the preprocessed sensor data into a preset database.

3. The linear guide fault diagnosis method based on deep learning according to claim 1, characterized in that, The first fault assessment result includes the probability value of ball wear fault, the probability value of guide rail corrosion fault, the probability value of slider deformation fault, and the probability value of slider end cover detachment fault.

4. The method for diagnosing linear guide faults based on deep learning according to claim 1, wherein, The method for constructing the guide rail fault assessment model includes: Collecting the operating state data of historical linear guides and preprocessing the collected data; Selecting a deep learning model as the basic architecture of the guide rail fault assessment model; the deep learning model includes a recurrent neural network and a convolutional neural network; Dividing the preprocessed operating state data of historical linear guides into a training set, a validation set, and a test set; Using the training set to train the model and adjusting the parameters of the model; Using the validation set data to validate the model and adjusting the parameters and structure of the model according to the validation results; Evaluating the model using the test set data; Deploying the trained model to the production environment, receiving the real-time data stream from the sensors, and performing online fault assessment.

5. The method for diagnosing faults of linear guide rails based on deep learning according to claim 1, wherein The mathematical calculation formula for correcting the first fault assessment result according to at least one set of the second fault assessment results is: Among them, P c represents the corrected fault probability value; P t represents the original target linear guide fault probability value; α is an adjustment coefficient between 0 and 1, used to balance the proportion of the original evaluation and the influence of homologous faults; n represents the number of associated linear guides; w i represents the weight of the i-th associated linear guide; P ai represents the fault probability value of the i-th associated linear guide.

6. The method for diagnosing linear guide faults based on deep learning according to claim 1, wherein, The fault diagnosis result of the target linear guide includes the corrected fault probability value, fault type confirmation, fault severity assessment, homologous fault influence analysis, and maintenance suggestions.

7. A linear guide fault diagnosis system based on deep learning, characterized in that, The system is applied to the deep learning-based linear guide fault diagnosis method as described in claim 1, and the system includes: A data collection module that obtains the first operating state data of the target linear guide and the second operating state data of at least one associated linear guide; A guide rail fault assessment module that inputs the first operating state data into a preset guide rail fault assessment model to obtain a first fault assessment result; The associated guide rail fault assessment module inputs at least one set of the second operating state data into the guide rail fault assessment model respectively to obtain at least one set of second fault assessment results; The fault diagnosis module corrects the first fault assessment result according to at least one set of the second fault assessment results considering the influence of homologous faults to obtain the target linear guide rail fault diagnosis result.

8. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected through the bus, and characterized in that, When the computer program is executed by the processor, it implements the steps in the method according to any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps in the method according to any one of claims 1-6.

Images