Bakelite guide rail fault monitoring method and system

Abstract

The invention provides a bakelite guide rail fault monitoring method and system, and relates to the technical field of fault detection, and the method comprises the steps: building a digital connection between a monitoring platform and a distributed sensor network; collecting multi-dimensional physical quantities during operation of the guide rail, and constructing a state information base; generating an abnormal state information base; calling the visual sensor group to collect a bakelite guide rail fault section image according to the abnormal sequence value, and obtaining a guide rail part image set; analyzing the guide rail part image set to obtain a guide rail fault cause library; performing correlation analysis according to the abnormal state information base and the guide rail fault cause base to obtain a guide rail fault model; and generating a fault maintenance scheme according to the guide rail fault model, and feeding back the fault maintenance scheme to the monitoring platform. The technical problems that in the prior art, a monitoring mode is not accurate and comprehensive enough, data processing capacity is lacked, judgment is slow and errors are prone to being generated due to the fact that a system mature fault model is lacked are solved.

Description

Technical Field

[0001] The present application relates to the technical field of fault detection, and particularly relates to a method and system for monitoring the faults of bakelite guide rails. Background Art

[0002] Bakelite guide rails play an indispensable role in many production fields. They provide accurate guidance and support for the operation of production equipment, and can ensure that each component moves stably along a predetermined trajectory. For example, in the furniture production field, the accuracy and stable operation of bakelite guide rails directly determine the cutting quality of solid wood boards. Any slight deviation of the guide rail may result in non-compliant cutting dimensions, affecting the overall quality of the furniture. Therefore, the normal operation of bakelite guide rails is directly related to the continuity of production and the stability of product quality. With the development of modern industrial production towards high precision, high efficiency, and large scale, the application scenarios of bakelite guide rails are constantly expanding, and their safe and stable operation has become one of the key factors to ensure the normal operation of the entire production system. Once a bakelite guide rail fails, it may cause the equipment to jam, increase wear, and even cause the entire production line to stop, resulting in huge economic losses. Moreover, a faulty bakelite guide rail may also cause damage to other related components, increasing the equipment maintenance cost and difficulty. Therefore, fault monitoring of bakelite guide rails is an inevitable requirement to ensure the smooth progress of production, reduce costs, and improve production efficiency.

[0003] Currently, there are many deficiencies in the monitoring technologies for bakelite guide rails. Most traditional monitoring technologies are based on regular manual inspections, which lack real-time performance. Staff can only check the guide rails at fixed time intervals, and it is difficult to detect problems at the first moment of a fault. Moreover, manual inspections are limited by the experience and skill levels of the staff, and there may be significant differences in the inspection results of different personnel, making it difficult to guarantee the accuracy of the detection results. Although some sensor-based monitoring technologies can achieve a certain degree of automated monitoring, the arrangement of sensors is often not reasonable enough, and they can only monitor partial areas of the guide rail, unable to comprehensively cover the entire operating state of the guide rail. In addition, the existing sensor monitoring technologies have slow judgment because there is no systematic and mature fault model. When facing the fault judgment of bakelite guide rails, there is a lack of comprehensive and accurate reference basis. This makes the diagnosis process not only time-consuming but also prone to errors.

[0004] In summary, the existing fault monitoring technologies for bakelite guide rails far from meet the industry requirements. Their monitoring methods are not precise and comprehensive enough, lack data processing capabilities, and the lack of a systematic and mature fault model leads to slow judgment and is prone to errors. Therefore, a method is needed to solve the above problems. Summary of the Invention

[0005] The present disclosure provides a method and system for monitoring the faults of bakelite guide rails, aiming to solve the technical problems in the prior art that the monitoring method is not accurate and comprehensive enough, lacks data processing capabilities, and lacks a systematic and mature fault model, resulting in slow judgment and easy generation of errors.

[0006] According to the first aspect of the present disclosure, there is provided a method for monitoring the faults of bakelite guide rails, including: Establish a digital connection between the monitoring platform and the distributed sensor network. The distributed sensor network includes M groups of multi-source sensors and M groups of visual sensor groups. The multi-source sensors are deployed at equal intervals on the edge of the bakelite guide rail, and the visual sensor groups are arranged around the guide rail in a surrounding manner according to the installation nodes of the multi-source sensors. Among them, the multi-source sensors include speed sensors, vibration sensors, and current sensors; Continuously collect multi-dimensional physical quantities during the operation of the guide rail through the multi-source sensors, and construct a status information library. The status information library includes speed information, vibration information, current information, and sequence values; Based on the anomaly detection algorithm, judge and extract the abnormal data in the status information library to generate an abnormal status information library. The abnormal status information library includes abnormal speed data sets, abnormal vibration data sets, abnormal current data sets, and abnormal sequence values; According to the abnormal sequence value, call the visual sensor group to collect images of the fault section of the bakelite guide rail to obtain a set of guide rail part images; Analyze the set of guide rail part images to obtain a guide rail fault cause library. The guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail loss fault cause set; According to the abnormal status information library and the guide rail fault cause library, perform a correlation analysis to obtain a guide rail fault model; Generate a fault repair plan according to the guide rail fault model, and feedback the fault repair plan to the monitoring platform.

[0007] According to the second aspect of the present disclosure, there is provided a system for monitoring the faults of bakelite guide rails, including: A sensor layout module, which is used to establish a digital connection between the monitoring platform and the distributed sensor network. The distributed sensor network includes M groups of multi-source sensors and M groups of visual sensor groups. The multi-source sensors are deployed at equal intervals on the edge of the bakelite guide rail, and the visual sensor groups are arranged around the guide rail in a surrounding manner according to the installation nodes of the multi-source sensors. Among them, the multi-source sensors include speed sensors, vibration sensors, and current sensors; A status information library construction module, which is used to continuously collect multi-dimensional physical quantities during the operation of the guide rail through the multi-source sensors, and construct a status information library. The status information library includes speed information, vibration information, current information, and sequence values; An abnormal state information library generation module, which is used to judge and extract abnormal data in the state information library based on an abnormal detection algorithm, and generate an abnormal state information library. The abnormal state information library includes an abnormal speed data set, an abnormal vibration data set, an abnormal current data set, and an abnormal sequence value; A fault image acquisition module, which is used to call the vision sensor group to acquire images of the faulty section of the bakelite guide rail according to the abnormal sequence value, and obtain a guide rail part image set; A fault cause analysis module, which is used to analyze the guide rail part image set to obtain a guide rail fault cause library. The guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail loss fault cause set; A fault model construction module, which is used to perform a correlation analysis based on the abnormal state information library and the guide rail fault cause library to obtain a guide rail fault model; A fault repair plan generation and feedback module, which is used to generate a fault repair plan based on the guide rail fault model and feedback the fault repair plan to the monitoring platform.

[0008] One or more technical solutions provided in the present disclosure have at least the following technical effects or advantages: Establish a digital connection between the monitoring platform and the distributed sensor network. The distributed sensor network includes M groups of multi-source sensors and M groups of visual sensor groups. The multi-source sensors are deployed at equal intervals on the edge of the bakelite guide rail, and the visual sensor groups are arranged around the guide rail according to the installation nodes of the multi-source sensors. Among them, the multi-source sensors include a speed sensor, a vibration sensor, and a current sensor. Continuously collect multi-dimensional physical quantities during the operation of the guide rail through the multi-source sensors, and construct a status information library. The status information library includes speed information, vibration information, current information, and sequence values. Based on the anomaly detection algorithm, judge and extract the abnormal data in the status information library to generate an abnormal status information library. The abnormal status information library includes an abnormal speed data set, an abnormal vibration data set, an abnormal current data set, and an abnormal sequence value. According to the abnormal sequence value, call the visual sensor group to collect images of the faulty section of the bakelite guide rail to obtain a set of guide rail part images. Analyze the set of guide rail part images to obtain a guide rail fault cause library. The guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail loss fault cause set. According to the abnormal status information library and the guide rail fault cause library, perform a correlation analysis to obtain a guide rail fault model. Generate a fault repair plan based on the guide rail fault model and feedback the fault repair plan to the monitoring platform. This solves the technical problems in the prior art that the monitoring method is not accurate and comprehensive enough, lacks data processing capabilities, and lacks a systematic and mature fault model, resulting in slow judgment and easy errors. It achieves the technical effects of accurate and rapid fault diagnosis, comprehensive monitoring, in-depth data analysis, and improvement of production safety and stability.

[0009] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of this application more obvious and understandable, the following specifically gives the specific implementation manners of this application. Brief Description of the Drawings

[0010] In order to more clearly illustrate the technical solutions in the present disclosure or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings described below are only exemplary, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0011] Figure 1 It is a schematic flowchart of a method for monitoring the faults of a bakelite guide rail provided by an embodiment of this application; Figure 2 It is a schematic structural diagram of a system for monitoring the faults of a bakelite guide rail provided by an embodiment of this application.

[0012] Description of the attached drawing reference numerals: Sensor layout module 11, status information library construction module 12, abnormal status information library generation module 13, fault image acquisition module 14, fault cause analysis module 15, fault model construction module 16, fault repair plan generation and feedback module 17. Detailed implementation manners

[0013] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0014] Embodiment 1. A method for monitoring the faults of bakelite guide rails provided by an embodiment of the present disclosure is hereinafter described with reference to Figure 1 as follows. The method includes: S1: Establish a digital connection between the monitoring platform and the distributed sensor network. The distributed sensor network includes M groups of multi-source sensors and M groups of vision sensor groups. The multi-source sensors are deployed at equal intervals on the edge of the bakelite guide rail, and the vision sensor groups are arranged around the guide rail according to the installation nodes of the multi-source sensors. Among them, the multi-source sensors include speed sensors, vibration sensors, and current sensors; Specifically, implement a segmented equal-distance deployment strategy on the mechanical structure of the guide rail. A group of multi-source sensors is deployed at every fixed distance, with a total of M groups. For example, determine the fixed distance value to be 1 meter, install a group of multi-source sensors at every 1 meter on the edge of the guide rail, and number them from 1 to M in sequence. Each group of multi-source sensors integrates three types of monitoring devices: speed sensors, vibration sensors, and current sensors. Each group of sensors is connected to the nearest edge computing node wirelessly to form a distributed data acquisition network, ensuring continuous state perception of the entire length of the guide rail. The deployment meets the requirements of equal-distance coverage, signal stability, and maintenance convenience. Arrange the vision sensor groups around the guide rail according to the installation nodes of the M groups of multi-source sensors and number them from 1 to M in sequence. Each group of vision sensor groups includes vision sensors in four directions, which are used to capture image information of the bakelite guide rail from different perspectives, realize visual coverage of the guide rail from all angles, obtain comprehensive image information of the guide rail, and provide rich visual data for subsequent fault analysis.

[0015] S2: Continuously collect multi-dimensional physical quantities during the operation of the guide rail through the multi-source sensors, and construct a status information library. The status information library includes speed information, vibration information, current information, and sequence values; Specifically, a non-contact photoelectric encoder is used as the speed sensor, a three-axis MEMS accelerometer is selected as the vibration sensor, and a closed-loop Hall current sensor is used as the current sensor. The time bases of multi-source data are aligned through a high-precision time synchronization protocol, the original signals are collected according to a preset sampling strategy, and preprocessing operations such as signal processing are performed and integrated and stored in the state information database. The state information database adopts a multi-dimensional matrix structure, with the multi-source sensor numbers as row indexes and the three types of monitoring data as column fields to construct a standardized data storage framework.

[0016] S3: Based on the anomaly detection algorithm, judge and extract the abnormal data in the state information database to generate an abnormal state information database, which includes an abnormal speed data set, an abnormal vibration data set, an abnormal current data set, and an abnormal sequence value; Specifically, deeply mine and analyze all the data in the state information database, and through three anomaly detection frameworks, accurately strip the abnormal data from the three types of monitoring data to construct an abnormal state information database. For speed information, use the Isolation Forest algorithm to identify abnormal patterns such as speed mutations, speed steady-state offsets, and excessive speed volatility, and extract the speed data that meets the abnormal judgment criteria, including sequence values, speed values, and speed abnormal types, to generate an abnormal speed data set; obtain the vibration information in the state information database, and use the K-Means clustering algorithm to screen abnormal vibration data and judge the abnormal type, and integrate the sequence value, vibration value, and abnormal vibration type into the abnormal vibration data set; obtain multi-position current data samples during the operation of the guide rail from the state information database. According to the current data under the normal operation state, draw a box plot, and compare the current value with the box plot threshold one by one. If the current value is less than the lower threshold or greater than the upper threshold, determine that the current information is abnormal. For the determined abnormal current data, construct an abnormal current data set, and each record in this data set should include a sequence value, a current value, and a current abnormal type. Integrate the abnormal speed data set, the abnormal vibration data set, and the abnormal current data set into the abnormal state information database with the sequence value as the reference index. Each sequence value may correspond to one or more abnormal data points, and these data points may come from different data sets. For example, an abnormal speed data and an abnormal current data may be included under one sequence value, indicating that both the speed and the current are abnormal under this sequence value number, or only abnormal vibration data may be included under one sequence value, indicating that only the vibration is abnormal under this specific sequence value number.

[0017] S4: Call the vision sensor group to collect images of the faulty section of the bakelite guide rail according to the abnormal sequence value to obtain a guide rail part image set; Specifically, according to the abnormal sequence value in the abnormal state information library, trigger the visual sensor group with the same number to capture the guide rail image. The visual sensor group collects images according to preset parameters such as resolution and shooting frequency to obtain the guide rail part image. According to the propagation speed of the traveling wave in the guide rail and the time difference of the traveling wave reaching different detection points, use the traveling wave positioning formula to calculate the specific position of the current fault in the guide rail. Use the LabelImg annotation tool to accurately mark the corresponding position of the current fault in the image on the guide rail part image with a rectangle or other appropriate marking shape, and generate a set of guide rail part images.

[0018] S5: Analyze the set of guide rail part images to obtain a guide rail fault cause library, where the guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail loss fault cause set; Specifically, for each guide rail part image, use the shape detection algorithm in image analysis technology to extract features. For example, adopt an algorithm based on edge detection, such as the Canny edge detection algorithm, to determine the edge contour of the guide rail part in the image. According to the edge contour, further extract shape features. Summarize the shape features extracted from all part images to generate a guide rail shape feature set. Use the texture analysis algorithm to extract texture features and construct a guide rail texture feature set. Select a suitable deformation measurement algorithm, such as the SIFT algorithm, and compare the known standard shape features with the currently collected shape features to calculate the deformation parameters compared with the standard shape, such as the curvature and compression rate of a certain part. According to the deformation parameters, analyze the possible causes of deformation. Summarize all possible causes to form a guide rail deformation fault cause set. Use the gray-level co-occurrence matrix and wavelet energy analysis method to decompose the guide rail texture features, calculate image contrast, energy, energy ratio parameters, and obtain the guide rail loss parameters. According to the loss parameters, analyze the wear state of the guide rail and obtain the loss causes. When analyzing the wear state, focus on positioning the marked position of the LabelImg tool to judge the loss state of the guide rail insulation layer. Add the guide rail loss causes to the guide rail loss fault cause set.

[0019] S6: According to the abnormal state information library and the guide rail fault cause library, conduct a correlation analysis to obtain a guide rail fault model; Specifically, extract the relevant data from the abnormal information library and the guide rail fault cause library, and align all the data with the sequence value as the index. Select a correlation measurement method, such as the Spearman rank correlation coefficient, to calculate the correlation between the abnormal state and the fault cause, construct a correlation matrix, where the rows of the matrix represent the abnormal state information, the columns of the matrix represent the fault cause set, and the elements in the matrix are the corresponding correlation coefficients. Analyze the correlation matrix to determine the degree of association between each abnormal state and the fault, and establish a guide rail fault model.

[0020] S7: Generate a fault repair plan based on the guide rail fault model and feedback the fault repair plan to the monitoring platform.

[0021] Specifically, generate a fault repair plan according to the fault classification in the guide rail fault model. For example, if it is determined that the guide rail has a thermal deformation fault, a maintenance strategy of adjusting the pre-tightening force of the guide rail installation bolts can be formulated. By appropriately adjusting the tightness of the bolts and utilizing the elasticity of the guide rail itself and the characteristics of the connection structure, an attempt is made to restore the flatness of the guide rail. Or use professional calibration tools, such as a guide rail straightening machine. The maintenance personnel need to first accurately measure the deformed part to determine the deformation amount and deformation direction, and then gradually perform the straightening operation on the guide rail according to the operation specifications of the straightening machine. Organize the generated fault repair plan in the format specified by the monitoring platform. For example, adopt a structured data format, including fields such as fault type, fault location, maintenance strategy, maintenance resource requirements, maintenance personnel requirements, and estimated maintenance time. Transmit the formatted fault repair plan data to the monitoring platform.

[0022] Further, step S2 of the present application further includes: Continuously collect the speed information, vibration information, and current information during the operation of the guide rail through the multi-source sensors, and adopt the IEEE 1588 Precision Time Protocol to ensure that the multi-parameter sampling time synchronization error is less than the set threshold; Add the speed information, vibration information, and current information collected by the same group of sensors to the same state information group, perform sequence value marking according to the multi-source sensor serial number, and integrate multiple state information groups to construct a state information library of the bakelite guide rail.

[0023] Specifically, each group of multi-source sensors integrates three types of monitoring devices: a speed sensor, a vibration sensor, and a current sensor. The speed sensor is a non-contact photoelectric encoder, the vibration sensor is a three-axis MEMS accelerometer, and the current sensor is a closed-loop Hall current sensor. At the location where the edge computing node is deployed, wireless communication configuration is performed on the sensors to ensure that each group of multi-source sensors can establish a stable communication connection with the nearest edge computing node, thereby forming a distributed data acquisition network. The formation of this network can ensure continuous state perception of the entire length of the guide rail, covering the entire working range of the guide rail and avoiding perception blind spots. A multi-source sensor time synchronization mechanism is set up, and the IEEE 1588 Precision Time Protocol is used to align the time reference of multi-source data. Through this protocol, it is ensured that the data collected by different sensors is highly synchronized in time. The IEEE 1588 Precision Time Protocol (PTP) is a protocol used to achieve high-precision time synchronization in Ethernet. Its working principle is to establish a master-slave hierarchy and achieve time and frequency synchronization by exchanging various messages. The sensor devices are debugged to receive the time synchronization signal from the master clock source, and their respective sampling clocks are adjusted according to this signal, so that the time error in the acquisition of speed, vibration, and current data is less than 1 μs. When the bakelite guide rail is running, data is collected at a preset sampling frequency to obtain the original speed information, vibration information, and current information. Operations such as filtering, noise reduction, and amplification are performed on the collected original information. For example, for the speed information collected by the speed sensor, a low-pass filter is used to remove high-frequency noise to make the information smoother and accurately reflect the true speed of the guide rail. For the data collected by the sensors in the same group, the speed information collected by the speed sensor, the vibration information collected by the vibration sensor, and the current information collected by the current sensor are integrated into the same state information group. Numbering is carried out according to the deployment order of the multi-source sensors on the guide rail, such as the 1st group, the 2nd group, etc., and an ordinal value label is given to each state information group. Multiple state information groups with ordinal value labels are integrated and integrated to construct a state information library for the bakelite guide rail. The state information library adopts a multi-dimensional matrix structure, with the ordinal value label of the sensor unit as the row index and the three types of monitoring data of speed, vibration, and current as the column fields, forming a standardized data storage framework. For example, if there are 10 groups of multi-source sensors, the matrix of the state information library will have 10 rows and 3 columns.

[0024] Further, step S3 of the present application further includes: Processing the speed information in the state information library using the Isolation Forest algorithm to extract abnormal speed data deviating from the normal mode and generating an abnormal speed data set; Selecting abnormal vibration data in the state information library based on the K-Means clustering algorithm to generate an abnormal vibration data set; Using the box plot method to define the guide rail current threshold in the normal operating state, screening the abnormal current data in the state information library, and generating an abnormal current data set; Obtain the sequence values corresponding to each abnormal data, link the abnormal data with the same sequence value, and integrate the abnormal speed dataset, abnormal vibration dataset, and abnormal current dataset into an abnormal state information database.

[0025] Specifically, the speed information in the state information database is processed based on the Isolation Forest algorithm. The algorithm performs multiple random splitting operations on the distribution of the speed information to form multiple isolation trees. Whether the data is abnormal speed data is determined according to the path length of the data points in the isolation tree. The speed abnormal types are divided according to the marked abnormal speed data. Analyze the change of the speed value in the sequence. Determine whether a mutation occurs by calculating the difference between adjacent speed values. If the absolute value of the speed derivative exceeds 0.25 m / s², the speed abnormal type is defined as a speed mutation type abnormality. Take the speed mean value within a long time window as the steady-state reference value, and calculate the speed mean value within a short time window. When the difference between the two mean values exceeds the preset threshold of 0.05 m / s, the relevant abnormal speed data is classified as a speed steady-state offset type abnormality. Calculate the volatility of the speed value, set the speed volatility threshold to 3% of the speed mean value. If the volatility exceeds this threshold, it is considered that the speed volatility exceeds the limit, and the corresponding speed data is marked as a speed volatility over-limit abnormality. Finally, all the speed data marked as abnormal is extracted in the structure of sequence value, speed value, and speed abnormal type to form an abnormal speed dataset.

[0026] Specifically, the K-Means clustering algorithm is used to analyze and process the vibration information captured by the vibration sensor. Based on the structural characteristics and operating mode of the guide rail, the initial clustering center of the K-Means clustering algorithm is determined. According to the K-Means clustering result, the vibration frequency far from the normal clustering center is determined as abnormal vibration data. The vibration abnormal types are divided according to the marked abnormal vibration data. Analyze the frequency value of the vibration information. If the main frequency > 500 Hz, the abnormal vibration type is judged as a high-frequency impact abnormality. If the main frequency < 50 Hz, the abnormal vibration type is judged as an unbalanced vibration abnormality. The vibration data determined to be abnormal, including the sequence value, frequency value, and vibration abnormal type, forms an abnormal vibration dataset.

[0027] Specifically, draw a box plot, calculate the lower quartile, upper quartile, and interquartile range, and define the current threshold when the bakelite guide rail operates normally. Mark the current information that deviates from the standard threshold as abnormal current data, and at the same time judge the specific current abnormal type according to the current value. If the current value is lower than the specific threshold, the abnormal current type is judged as undercurrent abnormality. If the current value is higher than the specific threshold, the abnormal current type is judged as overcurrent abnormality. The current data that meets the abnormal judgment criteria, including the sequence value, current value, and current abnormal type, is summarized to generate an abnormal current dataset.

[0028] Integrate the abnormal data under the same sequence value through data association. For example, establish an index table, use the sequence value as the index, and associate the corresponding abnormal speed data, abnormal vibration data, and abnormal current data. Store the integrated abnormal speed data set, abnormal vibration data set, and abnormal current data set into a unified abnormal status information database.

[0029] Furthermore, step S4 of this application further includes: Call the visual sensor group with the same number according to the abnormal sequence value, obtain the image of the faulty section of the bakelite guide rail, determine the specific location of the current fault based on the traveling wave positioning method, use the LabelImg annotation tool to mark the position on the image, and collect the images into the guide rail part image set; Specifically, trigger the visual sensor group with the corresponding number based on the abnormal sequence value data in the abnormal status information database. After receiving the trigger instruction, the visual sensor group performs image acquisition according to the preset parameters to obtain the guide rail part image. When a current fault occurs in the guide rail, a traveling wave will be generated at the fault point and propagate in all directions. According to the arrival time of the traveling wave detected by the current sensors in the sensor group, obtain the time difference of the traveling wave arriving at different detection points, and combine the propagation speed of the traveling wave. Use the traveling wave positioning formula to calculate the specific location of the current fault in the guide rail. For example, for a simple two-detection-point case, if the traveling wave propagation speed is v, the arrival times of the traveling wave at the two detection points are t 1 and t 2 , and the distance between the two detection points is d, then the fault location x can be calculated by the formula . Use the LabelImg tool to process the guide rail part image, and accurately mark the current fault location on the guide rail part image with the selected marking shape. Integrate all the guide rail part images to generate the guide rail part image set.

[0030] Furthermore, step S5 of this application further includes: Process the guide rail part image set through image analysis technology, extract the features of the guide rail part image, and obtain the guide rail shape feature set and the guide rail texture feature set; Combine the deformation measurement algorithm to perform deformation analysis on the guide rail shape feature set, obtain the guide rail deformation parameters, and construct the guide rail deformation fault cause set; Combine the loss measurement method to perform loss analysis on the guide rail texture feature set, obtain the guide rail loss parameters, and construct the guide rail loss fault cause set; Align the guide rail deformation fault cause set and the guide rail loss fault cause set according to the sequence value, and construct the guide rail fault cause database.

[0031] Specifically, preprocess the image set of the guide rail parts, and perform operations such as grayscaling, noise reduction, and enhancement on the part images. For example, use the Gaussian filtering algorithm to remove noise in the image and improve the image clarity, enhance the contrast through histogram equalization to highlight the texture features of the guide rail surface for subsequent feature extraction. Adopt an edge detection algorithm to extract the shape features of the guide rail, such as the Canny edge detection algorithm. This algorithm determines the edge position by calculating the gradient of the image, thereby obtaining the contour shape of the guide rail, and then constructs a guide rail shape feature set. Use the gray-level co-occurrence matrix method to extract texture features. Generate a gray-level co-occurrence matrix by calculating the occurrence frequency of pixel pairs with different gray values in the image, and then construct a guide rail texture feature set. Process the image through the SIFT algorithm, establish an affine transformation model to solve the parameter offset, and at the same time select points in the coordinate system to calculate the curvature and compression ratio of the guide rail, and accordingly construct a guide rail deformation fault cause set. Obtain the wear degree of the guide rail through the gray-level co-occurrence matrix and wavelet energy analysis, and construct a guide rail wear fault cause set. Using the ordinal value as the row index, take the guide rail deformation fault cause set and the guide rail wear fault cause set as two columns respectively, and align them in the order of the ordinal value to form a guide rail fault cause library in a table-like structure for subsequent correlation analysis.

[0032] Furthermore, this application also includes: Extract feature points on the guide rail part image through the SIFT algorithm, establish an affine transformation model between the reference image and the detection image, solve the deformation parameter matrix, and calculate the local coordinate offset; Establish a Cartesian coordinate system with the image center as the origin, the guide rail extension direction as the x-axis, and the vertical direction as the y-axis. Select three feature points at equal intervals along the longitudinal direction of the guide rail to calculate the curvature of the guide rail. The formula is: ; where k is the curvature, x 1 and y 1 are the horizontal and vertical coordinates of feature point 1, x 2 and y 2 are the horizontal and vertical coordinates of feature point 2, x 3 and y 3 are the horizontal and vertical coordinates of feature point 3; Based on the selected feature points, continue to calculate the compression ratio of the guide rail. The formula is: ; where e is the compression ratio, and L 0 represents the initial straight-line length of the guide rail when there is no external force and no deformation occurs; Determine the guide rail deformation fault cause according to the guide rail curvature and compression ratio, integrate the deformation fault causes of multiple images, and assign ordinal values to construct a guide rail deformation fault cause set.

[0033] Specifically, the SIFT algorithm is used to extract feature points in the image of the guide rail parts. The SIFT algorithm can detect representative feature points in different scale spaces, and these feature points contain key information of the image, such as local shape and other features. Based on the extracted feature points, extreme points are detected through the Gaussian difference pyramid, and a 128-dimensional feature descriptor is generated. The corresponding relationship between the feature points of the reference image and the detected image is established to construct an affine transformation model. Affine transformation is a linear transformation that can describe transformation relationships such as translation, rotation, scaling, and shear between images. Through this model, the corresponding relationship between the feature points of the two images can be found. The deformation parameter matrix of the affine transformation model is solved, and this matrix contains various parameter information about image transformation, such as rotation angle, scaling ratio, etc. By solving the matrix, the offset of the local coordinates is further calculated, and this offset reflects the position change of a certain local area in the image between the reference image and the detected image. According to the offset, it is determined whether there is installation stress release or thermal deformation fault in the guide rail.

[0034] Taking the image center as the origin, a Cartesian coordinate system is established. It is stipulated that the direction of the guide rail extension is the x-axis, and the direction perpendicular to the guide rail extension is the y-axis. Along the longitudinal direction of the guide rail, three feature points P 1 (x 1 , y 1 )、P 2 (x 2 , y 2 )、P 3 (x 3 , y 3 ) are selected at equal intervals. The curvature of the guide rail is calculated through the selected feature points, and the formula is: ; where x and y are the horizontal and vertical coordinates of the three feature points. When K > 0.15m -1 , it is determined that the guide rail has abnormal bending, and the fault cause is defined as foundation settlement or eccentric load.

[0035] Based on the selected feature points, the compression rate e of the guide rail is calculated, and the formula is: ; where L 0 represents the initial straight length of the guide rail when there is no external force and no deformation, and the numerator represents the actual length of the guide rail after deformation. The principle of the formula is to measure the degree of compression of the guide rail by comparing the ratio of the difference between the initial length and the actual length of the guide rail to the initial length. When e > 0.3%, the fault cause of the guide rail is defined as material creep or impact load.

[0036] Integrate the deformation fault causes corresponding to each part image, and construct a guide rail deformation fault cause set numbered by sequence values.

[0037] Furthermore, this application also includes: Analyzing the image contrast and energy using the gray-level co-occurrence matrix to obtain the wear depth of the guide rail; Decomposing the texture features of the guide rail according to the wavelet energy analysis method to obtain the sub-band energy and sub-band energy ratio parameters; Determining the loss degree of the guide rail based on the wear depth, energy, sub-band energy, and sub-band energy ratio parameters of the guide rail, obtaining the cause of the loss fault of the guide rail, integrating the causes of the loss faults of multiple images, and assigning sequence values to construct a set of causes of the loss faults of the guide rail; Specifically, based on the previously constructed gray-level co-occurrence matrix, the contrast and energy are calculated, and the formulas are as follows: Contrast: ; where P(i,j) represents the value of the element (i,j) in the gray-level co-occurrence matrix, and n represents the number of gray levels. For every 0.1 increase in the contrast C, the wear depth increases by 12 μm. Locate the current fault position marked by the LabelImg tool. If the wear depth at this position is greater than 0.3 mm, it is determined that the insulation layer is damaged.

[0038] Energy: ; where P(i,j) represents the value of the element (i,j) in the gray-level co-occurrence matrix, and n represents the number of gray levels. If the energy E drops by more than 15%, it is determined that the lubrication of the guide rail fails.

[0039] Select a suitable wavelet basis function, such as the Daubechies wavelet, to perform wavelet decomposition on the guide rail texture image. Multilevel wavelet decomposition is used to decompose the image into a low-frequency sub-band and high-frequency sub-bands. Calculate the energy of each sub-band, and the formula is: ; where, I k (x,y) is the pixel value at the coordinate (x,y) in the k-th sub-band image. This formula indicates that all pixel points in the sub-band image need to be traversed, and the square values of each pixel value are accumulated to obtain the sub-band energy E k .

[0040] Divide the sum of the high-frequency sub-band energy by the low-frequency sub-band energy to obtain the energy ratio parameter. Assume that the high-frequency sub-bands include LH3, HL3, HH3, and the low-frequency sub-band is LL3. The frequency band energy ratio is: ; where, E LH3 , E HL3 , E HH3 , E LL3They represent the energies of sub-bands LH3, HL3, HH3, and LL3 respectively. If D > 18%, it is judged that the guide rail loss fault is abrasive wear. If D < 5%, it is judged that the guide rail loss fault is adhesive wear.

[0041] Integrate the causes of loss faults corresponding to the images of each part, and construct a set of causes of guide rail loss faults numbered by sequence values.

[0042] Furthermore, step S6 of this application further includes: Extract the data in the abnormal state information library and the guide rail fault cause library, and align the data according to the sequence values; Use the Spearman rank correlation coefficient to analyze the correlation between the abnormal state and the fault cause. The formula is as follows: ; where n is the sample size, R xi and R yi are the ranks of the i-th observation in variables X and Y respectively; Obtain the guide rail fault model according to the correlation analysis results.

[0043] Specifically, extract the abnormal speed data set, abnormal vibration data set, and abnormal current data set in the abnormal state information library, and the guide rail deformation fault cause set and the guide rail loss fault cause set in the guide rail fault cause library. Use the sequence value as the core alignment dimension to achieve precise matching of the data sets, ensuring that each abnormal state information data corresponds logically to the corresponding fault cause data. Arrange and combine the abnormal speed data set, abnormal vibration data set, and abnormal current data set with the guide rail deformation fault cause set and the guide rail loss fault cause set respectively to construct abnormal state-fault cause combination pairs and perform correlation calculations.

[0044] According to the calculated correlation coefficient r, the correlation degree between the abnormal state and the fault cause can be judged. If r is high and close to 1, it means that the change of the abnormal state has a strong correlation with the fault cause, which may be an important indicator of this abnormal state. If r is low and close to 0, it means that the correlation between the two is weak, and other factors may need to be further analyzed.

[0045] Based on the judgment of the correlation coefficient, the main fault causes leading to each abnormal state are obtained. For example, if the abnormal state of this node is a speed mutation type abnormality and a high-frequency impact abnormality, according to the correlation coefficient, it can be judged that the fault cause at this stage is installation stress release or thermal deformation fault. Use the ensemble learning algorithm to construct a guide rail fault model, which can quickly judge the possible fault types according to the input abnormal state information, greatly shortening the fault detection time.

[0046] Embodiment 2. Based on the same inventive concept as the bakelite guide rail fault monitoring method in the foregoing embodiment, the present application further provides a bakelite guide rail fault monitoring system. Please refer to the attached Figure 2 , the system includes: A sensor layout module 11, which is used to establish a digital connection between the monitoring platform and the distributed sensor network. The distributed sensor network includes M groups of multi-source sensors and M groups of visual sensor groups. The multi-source sensors are evenly deployed on the edge of the bakelite guide rail, and the visual sensor groups are arranged around the guide rail according to the installation nodes of the multi-source sensors. Among them, the multi-source sensors include a speed sensor, a vibration sensor, and a current sensor; A status information library construction module 12, which is used to continuously collect multi-dimensional physical quantities during the operation of the guide rail through the multi-source sensors and construct a status information library. The status information library includes speed information, vibration information, current information, and sequence values; An abnormal status information library generation module 13, which is used to judge and extract abnormal data in the status information library based on an anomaly detection algorithm and generate an abnormal status information library. The abnormal status information library includes an abnormal speed data set, an abnormal vibration data set, an abnormal current data set, and an abnormal sequence value; A fault image acquisition module 14, which is used to call the visual sensor group to acquire images of the fault section of the bakelite guide rail according to the abnormal sequence value and obtain a guide rail part image set; A fault cause analysis module 15, which is used to analyze the guide rail part image set and obtain a guide rail fault cause library. The guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail loss fault cause set; A fault model construction module 16, which is used to perform a correlation analysis based on the abnormal status information library and the guide rail fault cause library to obtain a guide rail fault model; A fault repair plan generation and feedback module 17, which is used to generate a fault repair plan according to the guide rail fault model and feedback the fault repair plan to the monitoring platform.

[0047] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0048] The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for monitoring a bakelite rail fault, characterized in that: The method comprises: Establishing a digital connection between the monitoring platform and a distributed sensor network, wherein the distributed sensor network includes M groups of multi-source sensors and M groups of visual sensor groups, wherein the multi-source sensors are deployed at equal intervals on the edge of the bakelite rail, and the visual sensor groups are arranged around the rail according to the multi-source sensor installation nodes, wherein the multi-source sensors include speed sensors, vibration sensors, and current sensors; The multi-dimensional physical quantities of the guide rail during operation are continuously collected by the multi-source sensor to construct a state information library, wherein the state information library includes speed information, vibration information, current information and sequence value; Based on an abnormality detection algorithm, the abnormal data in the state information library is judged and extracted to generate an abnormal state information library, wherein the abnormal state information library includes an abnormal speed data set, an abnormal vibration data set, an abnormal current data set and an abnormal sequence value; According to the abnormal sequence value, the visual sensor group is called to collect the image of the faulty section of the bakelite guide rail to obtain a guide rail part image set; Analyze the guide rail part image set to obtain a guide rail fault cause library, wherein the guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail wear fault cause set; Performing correlation analysis based on the abnormal state information library and the guide rail fault cause library to obtain a guide rail fault model; A fault repair plan is generated according to the guide rail fault model, and the fault repair plan is fed back to the monitoring platform.

2. A method for monitoring a bakelite rail fault as claimed in claim 1, characterized in that: Build a status information library, including: The multi-source sensor continuously collects speed information, vibration information and current information during the operation of the guide rail, and adopts the IEEE 1588 precision time protocol to ensure that the multi-parameter sampling time synchronization error is less than a set threshold; The speed information, vibration information and current information collected by the same group of sensors are added to the same state information group, and the sequence values ​​are marked according to the sequence numbers of the multi-source sensors. Multiple state information groups are integrated to build a state information library for the bakelite guide rail.

3. A method for monitoring a bakelite guide rail fault as claimed in claim 1, characterized in that: Generate an abnormal status information library, including: The isolation forest algorithm is used to process the speed information in the state information database, extract the abnormal speed data that deviates from the normal mode, and generate an abnormal speed data set; Based on K-Means clustering algorithm, the abnormal vibration data in the state information library is selected to generate an abnormal vibration data set; The box plot method is used to define the rail current threshold in normal operation state, filter the abnormal current data in the state information library, and generate an abnormal current data set; The sequence value corresponding to each abnormal data is obtained, the abnormal data with the same sequence value are linked, and the abnormal speed data set, abnormal vibration data set and abnormal current data set are integrated into the abnormal state information library.

4. A method for monitoring a bakelite rail fault as claimed in claim 1, characterized in that: Get a collection of rail parts images, including: According to the abnormal sequence value, the visual sensor group with the same number is called to obtain the image of the faulty section of the bakelite rail, the specific location of the current fault is determined based on the traveling wave positioning method, the LabelImg annotation tool is used to mark the position of the image, and the image is collected into a guide rail part image set.

5. A method for monitoring a bakelite guide rail fault as claimed in claim 1, characterized in that: Get the rail fault cause library, including: Process the guide rail part image set by image analysis technology, extract features from the guide rail part images, and obtain the guide rail shape feature set and the guide rail texture feature set; Combined with the deformation measurement algorithm, deformation analysis is performed on the guide rail shape feature set to obtain the guide rail deformation parameters and construct the guide rail deformation fault cause set; Combined with the loss measurement method, the guide rail texture feature set is analyzed for loss, the guide rail loss parameters are obtained, and the guide rail loss fault cause set is constructed; The guide rail deformation fault cause set and the guide rail loss fault cause set are aligned according to the sequence value to build a guide rail fault cause library.

6. A method for monitoring a bakelite rail fault as claimed in claim 5, characterized in that: Construct a set of causes of guide rail deformation failures, including: The SIFT algorithm is used to extract feature points on the guide rail part image, establish an affine transformation model between the reference image and the detection image, solve the deformation parameter matrix, and calculate the local coordinate offset; A Cartesian coordinate system is established with the center of the image as the origin. The extension direction of the guide rail is the x-axis, and the vertical direction is the y-axis. Three feature points are selected at equal intervals along the longitudinal direction of the guide rail to calculate the curvature of the guide rail. The formula is: ； Where k is the curvature, x1 and y1 are the horizontal and vertical coordinates of feature point 1, x2 and y2 are the horizontal and vertical coordinates of feature point 2, and x3 and y3 are the horizontal and vertical coordinates of feature point 3; The guide rail compression rate is calculated based on the selected feature points. The formula is: ； Where e is the compression rate, L0 represents the initial straight length of the guide rail when there is no external force and no deformation; The causes of guide rail deformation faults are determined according to the curvature and compression rate of the guide rail. The deformation fault causes of multiple images are integrated and assigned sequence values ​​to construct a guide rail deformation fault cause set.

7. A method for monitoring a bakelite rail fault as claimed in claim 5, characterized in that: Construct a set of rail wear fault causes, including: The gray-level co-occurrence matrix is ​​used to analyze the image contrast and energy to obtain the rail wear depth. Decomposing the rail texture features according to the wavelet energy analysis method to obtain sub-band energy and sub-band energy ratio parameters; The rail wear degree is determined according to the rail wear depth, energy, sub-band energy and sub-band energy ratio parameters, and the rail wear fault cause is obtained. The wear fault causes of multiple images are integrated and assigned sequence values ​​to construct the rail wear fault cause set.

8. A method for monitoring bakelite rail faults as claimed in claim 1, characterized in that: Obtain guide rail fault models, including: Extract data from the abnormal status information database and the guide rail fault cause database, and align the data according to the sequence value; The Spearman rank correlation coefficient is used to perform correlation analysis between abnormal status and fault causes. The formula is as follows: ； Where n is the sample size, R xi and R yi are the ranks of the i-th observation in variables X and Y respectively; The guide rail fault model is obtained based on the correlation analysis results.

9. A bakelite rail fault monitoring system, characterized in that: The system comprises: A sensor deployment module, the sensor deployment module is used to establish a digital connection between the monitoring platform and a distributed sensor network, wherein the distributed sensor network includes M groups of multi-source sensors and M groups of visual sensor groups, wherein the multi-source sensors are deployed at equal intervals on the edge of the bakelite rail, and the visual sensor groups are arranged around the rail according to the multi-source sensor installation nodes, wherein the multi-source sensors include speed sensors, vibration sensors, and current sensors; A state information library construction module, which is used to continuously collect multi-dimensional physical quantities of the guide rail during operation through the multi-source sensor to construct a state information library, wherein the state information library includes speed information, vibration information, current information and sequence value; An abnormal state information library generation module, the abnormal state information library generation module is used to judge and extract abnormal data in the state information library based on an abnormal detection algorithm, and generate an abnormal state information library, wherein the abnormal state information library includes an abnormal speed data set, an abnormal vibration data set, an abnormal current data set and an abnormal sequence value; A fault image acquisition module, the fault image acquisition module is used to call the visual sensor group to acquire the image of the fault section of the bakelite guide rail according to the abnormal sequence value, and obtain a guide rail part image set; A fault cause analysis module, which is used to analyze the guide rail part image set and obtain a guide rail fault cause library, wherein the guide rail fault cause library includes a guide rail deformation fault cause set and a guide rail wear fault cause set; A fault model building module, the fault model building module is used to perform correlation analysis based on the abnormal state information library and the guide rail fault cause library to obtain a guide rail fault model; The fault repair plan generation and feedback module is used to generate a fault repair plan based on the guide rail fault model and feed back the fault repair plan to the monitoring platform.

Images