Belt machine roller voiceprint temperature integrated online monitoring system and method

Abstract

The application relates to a tape machine roller soundprint temperature integrated online monitoring system and method. A distributed auscultation optical fiber is arranged on a rack of a tape machine, and two ends of the distributed auscultation optical fiber are connected with an optical fiber auscultation analysis host computer; the optical fiber auscultation analysis host computer is configured to emit a first laser beam with a pulse to the distributed auscultation optical fiber and receive scattered light returned from the distributed auscultation optical fiber; a distributed temperature measurement optical fiber is arranged on a roller shaft of the tape machine, and two ends of the distributed temperature measurement optical fiber are connected with an optical fiber temperature analysis host computer; the optical fiber temperature analysis host computer is configured to emit a second laser beam with a pulse to the distributed temperature measurement optical fiber and receive scattered light returned from the distributed temperature measurement optical fiber; and the optical fiber auscultation analysis host computer and the optical fiber temperature analysis host computer are connected with a monitoring host computer. Abnormal vibration sound signals of the roller are captured through a vibration monitoring technology, temperature monitoring is used to discover temperature abnormal conditions in time, and the running state of the roller is accurately monitored.

Description

Technical Field

[0001] This application relates to the field of belt conveyor equipment fault detection, and in particular to an integrated online monitoring system and method for the acoustic signature and temperature of belt conveyor idler rollers. Background Technology

[0002] Belt conveyors are widely used in material handling equipment, featuring simple structure and high efficiency. They transport materials from one location to another using flexible conveyor belts and are widely used in various industries such as mines, ports, and factories.

[0003] The main components of a belt conveyor include drive rollers, idler rollers, and idler rollers, with the idler rollers serving to support and guide the conveyor belt. However, in actual use, idler rollers often face damage due to excessive weight of goods or other factors. Once an idler roller malfunctions, it may fail to rotate normally, leading to a series of problems such as increased friction and abnormal vibration. This not only affects the operating efficiency of the belt conveyor but may also damage the conveyor belt, increasing maintenance costs and downtime.

[0004] Traditional fault detection methods often rely on manual inspections, which can lead to problems such as slow response and missed detections, failing to meet the high requirements of modern industry for equipment reliability and efficiency. Summary of the Invention

[0005] Therefore, it is necessary to provide an integrated online monitoring system and method for acoustic signature and temperature of conveyor belt idler rollers.

[0006] One embodiment of this application is an integrated online monitoring system for acoustic signature and temperature of conveyor rollers, which includes distributed auscultation optical fiber, distributed temperature measurement optical fiber, optical fiber auscultation analysis host, optical fiber temperature analysis host and monitoring host.

[0007] The distributed auscultation fiber is mounted on the frame of the tape machine, and both ends of the distributed auscultation fiber are connected to the fiber optic auscultation analysis host.

[0008] The fiber optic auscultation analysis host is configured to emit a first laser beam with pulses into the distributed auscultation fiber and receive scattered light returned from the distributed auscultation fiber, locate a specific roller area based on Rayleigh scattering, and convert it into a vibration signal;

[0009] The distributed temperature-measuring optical fiber is mounted on the idler roller shaft of the conveyor belt, and both ends of the distributed temperature-measuring optical fiber are connected to the optical fiber temperature analysis host.

[0010] The fiber optic temperature analysis host is configured to emit a second laser beam with pulses into the distributed temperature measurement fiber and receive scattered light returned from the distributed temperature measurement fiber, locate a specific idler roller area based on Raman scattering, and convert it into a temperature signal;

[0011] The fiber optic auscultation analysis host and the fiber optic temperature analysis host are respectively connected to the monitoring host. The monitoring host determines that the idler roller has malfunctioned based on the vibration signal and temperature signal, and issues an alarm signal.

[0012] The aforementioned integrated online monitoring system for the acoustic signature and temperature of conveyor belt idlers utilizes a combination of distributed optical fibers for auscultation, distributed optical fibers for temperature measurement, an optical fiber auscultation analysis host, an optical fiber temperature analysis host, and a monitoring host. On one hand, it not only captures abnormal vibration signals from the idlers promptly through efficient vibration monitoring technology but also ensures timely detection of temperature anomalies through temperature monitoring. This combination of multi-sensor fusion advantages facilitates comprehensive and accurate monitoring of the idler's operating status. On the other hand, because it combines optical fiber auscultation analysis with optical fiber temperature analysis, it is simple to implement and easy to upgrade, with a relatively small footprint. This allows for convenient installation without affecting the normal operation of the conveyor belt, thus facilitating equipment modification and process improvement. Furthermore, it enables comprehensive monitoring of the conveyor belt... Real-time monitoring and fault early warning of idler rollers can promptly and accurately detect problems when rollers are damaged or unable to rotate. Combined with maintenance work, this can effectively prevent belt conveyor downtime and material transport interruptions caused by idler roller failures, thereby avoiding safety accidents and production stoppages due to malfunctions. This effectively improves overall work efficiency and equipment reliability. On the other hand, through distributed fiber optic sensing technology, the vibration and temperature changes of idler rollers can be accurately monitored, and potential faults can be identified in a timely manner. This ensures comprehensive monitoring of the entire belt conveyor system, avoids blind spots of traditional manual inspections, and effectively reduces reliance on manual inspections. This reduces human resource input, improves inspection efficiency and production automation, and is especially suitable for the coal production and transportation sector.

[0013] As an example, the fiber optic auscultation analysis host is configured to locate a specific idler roller area based on Rayleigh scattering, convert it into a vibration signal, and transmit it to the monitoring host; the fiber optic temperature analysis host is configured to locate a specific idler roller area based on Raman scattering, convert it into a temperature signal, and transmit it to the monitoring host.

[0014] As an example, the monitoring host is configured to receive vibration and temperature signals to determine whether a malfunction has occurred in the idler roller operation, and to issue an alarm signal if a malfunction has occurred in the idler roller operation.

[0015] In some embodiments, the distributed auscultatory optical fiber is configured to be clipped onto the frame of the tape machine.

[0016] As an example, the distributed auscultatory optical fiber is configured to be snapped into a channel steel on the frame of the conveyor belt. Alternatively, the integrated online monitoring system for acoustic signature and temperature of the conveyor belt idler rollers also includes a channel steel configured to be welded to the frame of the conveyor belt, with the distributed auscultatory optical fiber snapped into the channel steel.

[0017] In some embodiments, the distributed temperature-measuring optical fiber is positioned at a bolt position or a slot position on the outside of the idler roller shaft of the conveyor belt.

[0018] In some embodiments, the integrated online monitoring system for acoustic signature and temperature of the conveyor roller further includes a clamp, in which the distributed temperature-measuring optical fiber is wound, and the clamp is configured to magnetically attach to the outside of the conveyor roller shaft.

[0019] In some embodiments, the clamp is configured to magnetically attach to a bolt position or slot position on the outside of the idler shaft of the conveyor belt.

[0020] In some embodiments, the fiber optic auscultation analysis host and the fiber optic temperature analysis host are integrated into one unit.

[0021] In some embodiments, a method for integrated online monitoring of acoustic signature and temperature of conveyor belt idler rollers includes the following steps:

[0022] A first laser beam with pulses is emitted, and the first laser beam is transmitted through distributed stethoscope fibers;

[0023] It receives scattered light returned from the distributed auscultation fiber, locates specific idler roller areas based on Rayleigh scattering, and converts it into vibration signals;

[0024] A second laser beam with pulses is emitted, and the second laser beam is transmitted in a distributed temperature-sensing optical fiber;

[0025] It receives scattered light returned from the distributed temperature-measuring optical fiber, locates specific idler roller areas based on Raman scattering, and converts it into a temperature signal;

[0026] The vibration and temperature signals are cleaned and fused to obtain fused features for analysis and training. Based on the training results, it is determined whether the idler roller has malfunctioned, and an alarm signal is issued when the idler roller malfunctions.

[0027] In some embodiments, when a roller malfunctions, the monitoring host sends an alarm signal to a preset client.

[0028] In some embodiments, the integrated online monitoring method for the acoustic signature and temperature of the conveyor belt idler roller further includes the step of: when the idler roller malfunctions, including the fault area and fault status of the idler roller in the alarm signal based on the vibration signal and temperature signal.

[0029] In some embodiments, vibration signals and temperature signals are cleaned and fused to obtain fused features for analysis and training. The steps include: normalizing and extracting features from vibration signals and temperature signals to obtain cleaned signals; fusing the cleaned signals using a multi-sensor fusion method to obtain fused features; and inputting the fused features into a convolutional neural network for analysis and training to obtain training results.

[0030] In some embodiments, data normalization and feature extraction include the following steps: preprocessing vibration signals and temperature signals, dividing them into training sets and test sets; extracting features from vibration signals and temperature signals in the training set, inputting them into the training model for training, and using the training results as cleaned signals; dividing the test set into a registration set and a verification set, extracting features from vibration signals and temperature signals in the registration set and verification set respectively, comparing and verifying them, and using the verification results as cleaned signals. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram of the module connection of an embodiment of the integrated online monitoring system for acoustic signature and temperature of conveyor rollers described in this application.

[0033] Figure 2 for Figure 1 The illustrated embodiment is a schematic diagram used to show the positional relationship with the conveyor belt and its idler rollers.

[0034] Figure 3 for Figure 2 The diagram shown is a partial enlarged view of the embodiment, used to illustrate the fiber arrangement of a single idler roller.

[0035] Figure 4 This is a flowchart illustrating an embodiment of the integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers described in this application.

[0036] Figure 5 This is a flowchart illustrating another embodiment of the integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers described in this application.

[0037] Figure 6 This is a flowchart illustrating another embodiment of the integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers described in this application.

[0038] Figure 7 for Figure 6 The diagram shows a multi-sensor fusion neural network training process according to the embodiment shown.

[0039] Reference numerals in the attached diagram: 1. Distributed temperature measurement fiber optic cable; 2. Distributed auscultation fiber optic cable; 3. Conveyor belt roller shaft; 4. Temperature measurement fixture; 5. Conveyor belt bracket; 6. Fiber optic temperature analysis host; 7. Fiber optic auscultation analysis host; 8. Monitoring host. Detailed Implementation

[0040] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the specific embodiments of this application are described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0041] It should be noted that when a component is referred to as being "fixed to" or "set on" another component, it can be directly on the other component or there may be an intermediate component. When a component is considered to be "connected to" another component, it can be directly connected to the other component or there may be an intermediate component present. The terms "vertical," "horizontal," "upper," "lower," "left," "right," and similar expressions used in this application's specification are for illustrative purposes only and do not represent the only possible implementation.

[0042] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0043] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature and the second feature are in indirect contact through an intermediate medium. Furthermore, "above," "over," and "on top" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0044] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and or" as used in this application includes any and all combinations of one or more of the associated listed items.

[0045] In one embodiment of this application, an integrated online monitoring system for acoustic signature and temperature of conveyor belt idler rollers includes a distributed auscultatory fiber optic cable, a distributed temperature-measuring fiber optic cable, a fiber optic auscultatory analysis host, a fiber optic temperature analysis host, and a monitoring host. The distributed auscultatory fiber optic cable is mounted on the frame of the conveyor belt, with both ends connected to the fiber optic auscultatory analysis host. The fiber optic auscultatory analysis host is configured to emit a first laser beam with pulses into the distributed auscultatory fiber optic cable and receive scattered light returned from the distributed auscultatory fiber optic cable, locate a specific idler roller area based on Rayleigh scattering, and convert it into a vibration signal. The distributed temperature-measuring fiber optic cable is mounted on the idler roller shaft of the conveyor belt, with both ends connected to the fiber optic temperature analysis host. The fiber optic temperature analysis host is configured to emit a second laser beam with pulses into the distributed temperature-measuring fiber optic cable and receive scattered light returned from the distributed temperature-measuring fiber optic cable, locate a specific idler roller area based on Raman scattering, and convert it into a temperature signal. The fiber optic auscultatory analysis host and the fiber optic temperature analysis host are respectively connected to the monitoring host. The monitoring host determines that a fault has occurred in the idler roller operation based on the vibration signal and the temperature signal, and issues an alarm signal.

[0046] This design, through the coordinated use of distributed auscultation fiber optic cables, distributed temperature measurement fiber optic cables, fiber optic auscultation analysis host, fiber optic temperature analysis host, and monitoring host, not only captures abnormal vibration signals of the idler rollers in a timely manner through efficient vibration monitoring technology, but also ensures timely detection of temperature anomalies through temperature monitoring. This combines the advantages of multi-sensor fusion, facilitating comprehensive and accurate monitoring of the idler roller's operating status. Furthermore, because it combines fiber optic auscultation analysis with fiber optic temperature analysis, it is simple to implement and easy to upgrade, with a relatively small footprint. This allows for convenient installation without affecting the normal operation of the conveyor belt, thus facilitating equipment modification and process improvement. Finally, it enables real-time monitoring of the conveyor belt idler rollers. The monitoring and fault early warning system can detect problems promptly and accurately when the idler rollers are damaged or unable to rotate. Combined with maintenance work, it can effectively prevent belt conveyor downtime and material transport interruptions caused by idler roller failures, thereby avoiding safety accidents and production stoppages caused by failures, and thus effectively improving overall work efficiency and equipment reliability. On the other hand, through distributed fiber optic sensing technology, it can accurately monitor the vibration and temperature changes of the idler rollers, identify potential faults in a timely manner, ensure comprehensive monitoring of the entire belt conveyor system, avoid blind spots of traditional manual inspections, and effectively reduce reliance on manual inspections, thereby reducing human resource input, improving inspection efficiency and production automation, which is especially suitable for the coal production and transportation fields.

[0047] To monitor abnormal vibrations in conveyor belts and promptly detect problems with idlers, especially by segment-by-segment detection to pinpoint the location of faulty idlers, distributed auscultatory optical fibers are mounted on the conveyor belt frame in various embodiments. In some embodiments, the distributed auscultatory optical fibers are snapped onto the conveyor belt frame. For example, the distributed auscultatory optical fibers are snapped onto the conveyor belt frame along the conveyor belt's transport direction, meaning the length of the distributed auscultatory optical fibers is parallel to the conveyor belt's transport direction and snapped onto the conveyor belt frame; alternatively, the distributed auscultatory optical fibers are tightly bound to the conveyor belt frame along the conveyor belt's transport direction using strapping. For example, the distributed auscultatory optical fibers are configured to be snapped into a channel steel on the conveyor belt frame. Alternatively, the integrated online monitoring system for the acoustic signature and temperature of conveyor idlers also includes a channel steel, which is configured to be welded to the conveyor belt frame, with the distributed auscultatory optical fibers snapped into the channel steel.

[0048] This design offers several advantages. First, the distributed auscultation fiber optic cables are directly clipped or bundled into the conveyor belt frame or its channel steel. This simple, flexible, and secure installation ensures a high degree of fit and allows for precise coupling of vibration acoustic signals generated by the idler rollers. This effectively reduces signal attenuation and external interference, guaranteeing the stability and reliability of vibration monitoring. Second, the distributed auscultation fiber optic cables are laid parallel to the conveyor belt's direction, enabling continuous, distributed vibration acquisition across the entire frame. This allows for precise capture of subtle vibration differences in each idler roller section, quickly pinpointing the specific idler roller location corresponding to abnormal vibrations. This completely solves the problems of traditional monitoring methods, which struggle with segment-by-segment troubleshooting and suffer from ambiguous location. This method significantly improves fault location efficiency. Furthermore, the installation method, where the fiber optic cable is clipped into the frame channel steel or a specially welded channel steel, effectively protects the distributed auscultation fiber, preventing damage from dust, material impact, and mechanical abrasion, thus extending equipment lifespan. Simultaneously, the installation process requires no major alterations to the original structure of the conveyor belt, ensuring normal equipment operation and adapting to both new and old equipment retrofit scenarios, reducing construction difficulty and retrofit costs. Moreover, this deployment method is suitable for complex working conditions such as coal production and transportation, enabling long-term stable operation, reducing the frequency of subsequent maintenance, further lowering operation and maintenance costs, and effectively ensuring the continuous and reliable operation of the monitoring system.

[0049] To monitor abnormal vibrations of the idler rollers, in various embodiments, both ends of the distributed auscultatory fiber are connected to a fiber optic auscultatory analysis host. The fiber optic auscultatory analysis host is configured to emit a pulsed first laser beam into the distributed auscultatory fiber and receive scattered light returned from the distributed auscultatory fiber. That is, the incident end of the distributed auscultatory fiber is connected to the fiber optic auscultatory analysis host, and the fiber optic auscultatory analysis host emits a pulsed first laser beam into the incident end of the distributed auscultatory fiber. This first laser beam can be simply referred to as the first laser to distinguish it from the second laser beam in other embodiments. Furthermore, the emitting end of the distributed auscultatory fiber is also connected to the fiber optic auscultatory analysis host, and the fiber optic auscultatory analysis host receives the scattered light returned from the emitting end of the distributed auscultatory fiber. In some embodiments, the fiber optic auscultatory analysis host is configured to locate a specific idler roller area based on Rayleigh scattering and convert it into a vibration signal. In some embodiments, the fiber optic auscultatory analysis host is configured to emit a pulsed first laser beam into the distributed auscultatory fiber and receive scattered light returned from the distributed auscultatory fiber, locate a specific idler roller area based on Rayleigh scattering, and convert it into a vibration signal.

[0050] This design achieves several advantages. First, it uses a fiber optic auscultation analysis host to transmit and receive pulsed first laser beams at both ends of a distributed auscultation fiber, forming a closed-loop signal transmission path. This efficiently excites and collects Rayleigh scattering signals within the fiber, ensuring stable signal transmission and rapid response, effectively improving the sensitivity and completeness of vibration signal capture. Second, based on the Rayleigh scattering principle, it precisely locates specific idler roller areas, directly converting the scattered light signal into a vibration signal. This enables real-time analysis and precise source tracing of vibration data, avoiding distortion and delay during signal conversion and ensuring the accuracy of monitoring results. Third, the dual-end connection simplifies the system architecture, reduces signal transfer links, lowers equipment failure rates, and allows for synchronous acquisition and analysis of signals across the entire fiber optic cable, achieving full-area coverage monitoring of idler roller vibration and eliminating blind spots. Fourth, the use of laser pulse and Rayleigh scattering sensing technology provides strong resistance to electromagnetic interference, making it suitable for harsh industrial environments such as coal transportation. It can operate stably for extended periods, providing reliable technical support for early identification and accurate warning of abnormal idler roller vibration, ensuring the safe and efficient operation of the conveyor belt system.

[0051] To monitor the temperature of the idler rollers and detect problems promptly, in various embodiments, a distributed temperature-sensing optical fiber is disposed on the idler roller shaft of the conveyor belt. As an example, the distributed temperature-sensing optical fiber is disposed on the outside of the idler roller shaft. Alternatively, because temperature measurement is required, the distributed temperature-sensing optical fiber may directly contact the idler roller shaft, indirectly contact the idler roller shaft through a heat-conducting device, or be disposed adjacent to the idler roller shaft so as to be affected by the temperature of the idler roller shaft. In some embodiments, the distributed temperature-sensing optical fiber is disposed at a bolt position or a slot position on the outside of the idler roller shaft of the conveyor belt.

[0052] This design offers several advantages. First, by directly or indirectly attaching the distributed temperature-sensing optical fiber to the outside of the idler shaft, or placing it in key locations such as bolts and slots, it efficiently conducts heat from the idler shaft, accurately senses temperature changes, and effectively improves the sensitivity and response speed of temperature monitoring. Second, the attached arrangement reduces environmental interference, ensuring the authenticity and stability of temperature signal acquisition and avoiding the problems of poor contact and inaccurate temperature measurement inherent in traditional monitoring methods, thus achieving continuous and real-time monitoring of the idler temperature. Third, the flexible deployment method adapts to idler shafts of different specifications, is easy to install without complex modifications, and does not affect the normal rotation of the idlers, making it suitable for both new and old conveyor belt equipment retrofitting scenarios. Fourth, distributed temperature measurement can cover all idlers on the entire machine, eliminating blind spots in single-point monitoring, promptly identifying potential faults such as overheating and jamming, reducing reliance on manual inspections, minimizing safety hazards, ensuring the long-term stable operation of the conveyor belt, and effectively improving equipment reliability and production efficiency.

[0053] To facilitate monitoring of idler roller temperature, in some embodiments, the integrated online monitoring system for the acoustic signature temperature of the conveyor roller also includes a clamp, with a distributed temperature-sensing optical fiber wound inside the clamp. For example, the distributed temperature-sensing optical fiber is wound around the clamp in a spiral shape with at least five turns. For example, the clamp is configured to magnetically adhere to the outside of the conveyor roller shaft or be axially connected to the outside of the conveyor roller shaft. In some embodiments, the clamp is configured to magnetically adhere to a bolt position or slot position on the outside of the conveyor roller shaft. For example, the clamp is a heat-conducting device and is magnetically adhered to the outside of the conveyor roller shaft in a semi-enclosed manner, used to conduct heat from the idler roller shaft to the distributed temperature-sensing optical fiber without affecting the rotation of the idler roller shaft.

[0054] This design achieves several advantages. First, by using a dedicated clamp to wind the distributed temperature-sensing optical fiber into a spiral shape with at least five turns, it significantly increases the contact area between the fiber and the clamp, enhancing heat transfer efficiency, improving temperature sensing sensitivity, and effectively avoiding the problems of inaccurate single-point contact temperature measurement and response lag. Second, the clamp is fixed to the outside of the idler roller shaft, bolt, or slot using magnetic or shaft connection methods, making installation convenient and disassembly flexible without altering the original structure of the idler roller, and without affecting its normal rotation, allowing for rapid deployment under different working conditions. Third, the clamp uses heat-conducting fixtures and is semi-enclosed to the idler roller shaft, which can efficiently conduct heat from the shaft to the optical fiber while protecting the fiber from dust and impact damage, extending its service life. Fourth, the spiral winding combined with the heat-conducting clamp design enables continuous and stable acquisition of the idler roller temperature, eliminating monitoring blind spots, providing timely warnings of overheating faults, reducing the intensity of manual inspections, ensuring the safe and reliable operation of the conveyor belt, and improving production efficiency and maintenance levels.

[0055] In various embodiments, both ends of the distributed temperature-sensing optical fiber are connected to a fiber optic temperature analysis host. The fiber optic temperature analysis host is configured to emit a pulsed second laser beam into the distributed temperature-sensing optical fiber and receive scattered light returned from the distributed temperature-sensing optical fiber. That is, the incident end of the distributed temperature-sensing optical fiber is connected to the fiber optic temperature analysis host, which emits a pulsed second laser beam into the incident end of the distributed temperature-sensing optical fiber. Furthermore, the emitting end of the distributed temperature-sensing optical fiber is also connected to the fiber optic temperature analysis host, which receives scattered light returned from the emitting end of the distributed temperature-sensing optical fiber. In some embodiments, the fiber optic temperature analysis host is configured to locate a specific idler roller area based on Raman scattering and convert it into a temperature signal. In some embodiments, the fiber optic temperature analysis host is configured to emit a pulsed second laser beam into the distributed temperature-sensing optical fiber and receive scattered light returned from the distributed temperature-sensing optical fiber, locate a specific idler roller area based on Raman scattering, and convert it into a temperature signal.

[0056] This design offers several advantages. First, the fiber optic temperature analysis host emits and receives pulsed second laser beams at both ends of the distributed temperature-measuring fiber, forming a complete closed-loop optical path. This allows for stable excitation and acquisition of Raman scattering signals within the fiber, providing strong anti-interference capabilities and low transmission loss, effectively ensuring the continuity and reliability of temperature signal acquisition. Second, based on the Raman scattering principle, it precisely locates specific idler roller areas, directly converting the scattered light signal into a temperature signal. This enables real-time analysis and precise tracing of temperature data, avoiding signal conversion distortion and significantly improving temperature measurement accuracy and response speed. Third, the two-end connection architecture simplifies system composition, reduces signal transfer links, lowers the probability of failure, and allows for synchronous acquisition and analysis of the entire fiber optic section, achieving full-area temperature coverage monitoring of all idler rollers and eliminating monitoring blind spots. Fourth, it effectively adapts to complex working conditions such as coal production and transportation, ensuring long-term stable operation and providing early warnings for idler roller overheating and jamming. This effectively reduces reliance on manual inspections, ensuring the safe and efficient operation of the conveyor belt, thereby improving equipment reliability and production continuity.

[0057] To collect signals for confirming whether the idler rollers of the conveyor belt have malfunctioned, in various embodiments, the fiber optic auscultation analysis host and the fiber optic temperature analysis host are respectively connected to the monitoring host. In some embodiments, the monitoring host determines that the idler roller has malfunctioned based on vibration and temperature signals and issues an alarm signal; that is, the monitoring host determines whether the idler roller has malfunctioned based on vibration and temperature signals and issues an alarm signal when the idler roller has malfunctioned. As an example, the monitoring host is configured to determine whether the idler roller has malfunctioned based on signals from the fiber optic auscultation analysis host and the fiber optic temperature analysis host, and issues an alarm signal when the idler roller has malfunctioned. In some embodiments, the fiber optic auscultation analysis host and the fiber optic temperature analysis host are integrated. In some embodiments, the monitoring host, the fiber optic auscultation analysis host, and the fiber optic temperature analysis host are integrated.

[0058] This design offers several advantages. First, the fiber optic auscultation analysis host and the fiber optic temperature analysis host are respectively connected to the monitoring host, enabling unified collection and centralized processing of both vibration and temperature signals. This facilitates multi-source data fusion analysis, comprehensively assesses the roller's operating status, and effectively improves fault identification accuracy. Second, the monitoring host comprehensively analyzes the output signals from both hosts, promptly issuing alarm signals upon detecting anomalies, achieving rapid fault warning and shortening response time. Third, it supports integrated setup of multiple devices, simplifying system architecture, reducing equipment space requirements, lowering wiring complexity and hardware costs, and facilitating on-site installation, commissioning, and subsequent maintenance. Fourth, the integrated design boasts strong anti-interference capabilities, stable and reliable operation, and adaptability to complex industrial environments, effectively reducing manual intervention.

[0059] To facilitate confirmation of whether the idler rollers of the conveyor belt have malfunctioned, as an example, a fiber optic auscultation analysis host is configured to locate specific idler roller areas based on Rayleigh scattering, convert the signals into vibration signals, and transmit them to the monitoring host; a fiber optic temperature analysis host is configured to locate specific idler roller areas based on Raman scattering, convert the signals into temperature signals, and transmit them to the monitoring host. In this embodiment, different optimization designs are used for the detection of vibration and temperature signals. Rayleigh scattering is elastic scattering; when photons collide with matter molecules, no energy exchange occurs, so the frequency and wavelength of the scattered light are exactly the same as the incident light, only the direction of light propagation changes. Raman scattering, on the other hand, is inelastic scattering; when photons collide with matter molecules, energy exchange occurs, so the frequency and wavelength of the scattered light change slightly. Based on this, Rayleigh scattering is used to locate specific idler roller areas, such as the location of abnormal vibration, while Raman scattering is used to locate specific idler roller areas, such as the location of abnormally high temperatures. As an example, the monitoring host is configured to receive vibration and temperature signals to determine whether the idler roller operation has malfunctioned, and to issue an alarm signal when the idler roller operation malfunctions.

[0060] This design achieves several advantages. First, it optimizes vibration and temperature monitoring using differentiated scattering principles. Leveraging Rayleigh scattering's characteristics of no energy exchange and constant frequency wavelength, it accurately captures abnormal roller vibrations and locates the vibration area, resulting in strong signal stability and high positioning accuracy. Second, it utilizes Raman scattering's energy exchange and frequency wavelength changes to accurately sense roller temperature changes and locate high-temperature areas, offering high temperature sensitivity and rapid response. These two technologies complement each other to meet different monitoring needs. Third, the two types of signals are independently acquired and accurately traced before being uniformly transmitted to the monitoring host, enabling synchronous aggregation and fusion analysis of vibration and temperature data. This comprehensively depicts the roller's operating status, effectively avoiding the limitations of single-parameter monitoring and significantly improving the accuracy and reliability of fault identification. Fourth, the monitoring host comprehensively analyzes the two types of signals and promptly issues alarms, achieving accurate fault warnings and rapid fault location, reducing false alarms and missed alarms, lowering the intensity of manual inspections, ensuring the safe and stable operation of the conveyor belt, effectively preventing production interruptions and safety hazards caused by equipment failures, and improving overall operation and maintenance efficiency and production continuity.

[0061] To achieve real-time monitoring and fault early warning of conveyor belt idlers, one embodiment includes an integrated online monitoring system for conveyor belt idlers based on acoustic signature and temperature. This system aims to achieve real-time status assessment and fault early warning of conveyor belt idlers by combining acoustic signature and temperature monitoring. The system includes distributed auscultation fiber optic cables, distributed temperature measurement fiber optic cables, a fiber optic temperature analysis host, a fiber optic auscultation analysis host, and a monitoring host. The distributed auscultation fiber optic cables monitor the acoustic and vibration signals of the idlers in real time, while the distributed temperature measurement fiber optic cables monitor the temperature changes of the idlers in real time. The fiber optic temperature analysis host converts the optical signals into temperature signals and sends them to the monitoring host in real time. The fiber optic auscultation analysis host converts the optical signals into acoustic and vibration signals and sends them to the monitoring host. The monitoring host simultaneously receives signals from both analysis hosts and uses a multi-feature fusion neural network to calculate idler fault modes in real time. Based on the analysis results, it provides fault early warnings, achieving accurate diagnosis of the idler status and thus providing full-coverage monitoring of the entire conveyor belt line, especially the conveyor belt idlers.

[0062] As an example, the distributed auscultation fiber optic cable is installed within a channel steel welded to the outside of the conveyor belt frame using pre-set clamps. The distributed auscultation fiber optic cable terminal is connected to the fiber optic auscultation analysis host. The laser emitted by the light source of the fiber optic auscultation analysis host is transmitted through the distributed auscultation fiber optic cable, and Rayleigh scattering light information is fed back to the fiber optic auscultation analysis host. The fiber optic auscultation analysis host converts this into an acoustic vibration signal in real time and sends it to the monitoring host to obtain the current real-time idler roller fault mode.

[0063] As an example, a distributed temperature-sensing optical fiber is wound inside a pre-set clamp, which is magnetically fixed to the bolt position on the outside of the roller shaft. The distributed temperature-sensing optical fiber terminal is connected to the optical fiber temperature analysis host. The laser emitted by the light source of the optical fiber temperature analysis host is transmitted in the distributed temperature-sensing optical fiber and returns temperature change characteristic information through the Raman scattering principle. The optical fiber temperature analysis host collects data in real time, converts it into a temperature signal, and sends it to the monitoring host.

[0064] As an example, the monitoring host is connected to two analysis hosts via ports, including a fiber optic auscultation analysis host and a fiber optic temperature analysis host. The monitoring host is responsible for receiving and processing the acoustic vibration and temperature signals from these two analysis hosts. Through multi-sensor fusion neural network training, the monitoring host can accurately infer the fault mode of the idler roller. Once an idler roller fault is detected, the monitoring host immediately feeds back the fault information to the client, enabling maintenance personnel to carry out timely and effective maintenance based on the provided information.

[0065] This design, combining distributed fiber optic acoustic sensing technology with high-sensitivity temperature sensing fiber optics, effectively improves the accuracy and reliability of monitoring. By comprehensively analyzing acoustic and temperature signals, the working status of the conveyor belt idler rollers is monitored in real time. A sound signature feature library is established by modeling the sound signatures emitted by the idler rollers under normal operating conditions and comparing them with real-time acquired sound signature signals to promptly identify abnormal noise. Simultaneously, data from the temperature sensing fiber optics is used to monitor temperature changes in the idler rollers, further assessing the equipment's health status.

[0066] As an example, the monitoring host is equipped with a convolutional neural network, i.e., a convolutional neural network model, or a multi-sensor fusion convolutional neural network model. The convolutional neural network model is a multi-branch one-dimensional convolutional neural network structure, corresponding to three types of monitoring data collected synchronously from the same source: fiber optic vibration sensing signals, acoustic signature sensing signals, and temperature sensing signals. Each branch is set with an independent feature extraction branch. After each branch completes the adaptive extraction of shallow, middle, and deep fault features of single-mode sensing data, the deep fusion of multi-source heterogeneous sensing data is achieved through a feature-level fusion mechanism. This effectively makes up for the defects of single-sensor data features being one-sided and having weak anti-interference ability, matches the core technical requirements of multi-sensor fusion algorithms, and realizes accurate identification and judgment of faults such as roller wear, jamming, abnormal noise, and overheating. As an example, the multi-sensor fusion convolutional neural network model includes: three parallel single-modal feature extraction branches, a feature splicing and fusion module, a fully connected feature optimization module, and a fault classification output module; wherein, the first branch is the fiber optic vibration signal feature extraction branch, the second branch is the acoustic signature signal feature extraction branch, and the third branch is the temperature time series signal feature extraction branch. The outputs of the three feature extraction branches are all connected to the input of the feature splicing and fusion module, and the output of the feature splicing and fusion module is sequentially connected to the fully connected feature optimization module and the fault classification output module.

[0067] Furthermore, in the specific hierarchical structure of the convolutional neural network model, each single-modal feature extraction branch adopts a one-dimensional convolutional neural network structure. Each branch is composed of an input layer, a two-level convolutional pooling combination layer, a one-level deep convolutional layer, and a global average pooling layer stacked in sequence. The structural parameters of each branch are adapted to the temporal characteristics of the corresponding sensing signal. The specific hierarchical settings are as follows.

[0068] Input layer: used to receive pre-processed single-mode time-series sensing data, wherein the input dimension of fiber optic vibration signal is set to 1×1024, the input dimension of acoustic fingerprint signal is set to 1×1024, and the input dimension of temperature time-series signal is set to 1×256. All input data are single-channel normalized time-series sequences to ensure the standardization of multi-mode data input.

[0069] The first-level convolutional pooling layer consists of a first one-dimensional convolutional layer and a first max-pooling layer. The first one-dimensional convolutional layer has 32 convolutional kernels with a kernel size of 16×1 and a stride of 2. It uses Same padding and the ReLU activation function to extract shallow temporal features of the sensing signal, including signal amplitude abrupt changes, short-term impacts, and basic fluctuation features. The first max-pooling layer has a pooling kernel size of 4×1 and a stride of 4. It is used to reduce the dimensionality of the convolutional features, retain the core effective features, and remove redundant noise information.

[0070] The second-level convolutional pooling layer consists of a second one-dimensional convolutional layer and a second max pooling layer. The second one-dimensional convolutional layer has 64 convolutional kernels with a kernel size of 8×1 and a stride of 2. It uses Same padding and the ReLU activation function to mine mid-level correlation features in the signal, including vibration frequency features, acoustic spectrum features, and temperature temporal fluctuation patterns. The second max pooling layer has a pooling kernel size of 4×1 and a stride of 4, which further compresses the feature dimension and simplifies the model's computational load.

[0071] Deep convolutional layer: This is the third one-dimensional convolutional layer, with 128 convolutional kernels, a kernel size of 4×1, a stride of 2, and Same padding. The activation function is ReLU, which is used to extract deep latent features strongly correlated with idler roller faults, including bearing wear vibration features, idler roller jamming and abnormal noise features, and fault-corresponding temperature change features.

[0072] Global average pooling layer: connected to the output of deep convolutional layer, it converts one-dimensional feature map into one-dimensional feature vector with fixed dimension, abandons the traditional fully connected flattening method, effectively suppresses model overfitting, and achieves normalized output of single-modal features.

[0073] As an example, the monitoring host is equipped with a multi-sensor feature fusion and classification system, which includes: a feature splicing and fusion module, a fully connected feature optimization module, and a fault classification output module.

[0074] The feature splicing and fusion module splices and fuses the three independent one-dimensional feature vectors output from the fiber optic vibration feature branch, acoustic feature branch, and temperature feature branch in the channel dimension to form a global fused feature vector containing multi-dimensional information of vibration, acoustics, and temperature, thus completing the feature-level deep fusion of multi-sensor data.

[0075] The fully connected feature optimization module consists of two fully connected layers. The first fully connected layer has 256 neurons, which are used with the ReLU activation function and a Dropout regularization mechanism with a ratio of 0.3 to optimize the weights of the global fused features and remove redundant features. The second fully connected layer has 64 neurons to further refine the core fault features and improve the feature recognition rate.

[0076] The fault classification output module uses the Softmax activation function as the output layer activation function. The number of output neurons corresponds to the classification of the idler roller's operating status, specifically including five types of operating conditions: normal operation, bearing wear, idler roller jamming, abnormal noise, and local overheating. It outputs the probability value of each type of operating condition to achieve intelligent identification of the idler roller's operating status.

[0077] To achieve real-time monitoring and fault early warning of conveyor belt idler rollers, in some embodiments, an integrated online monitoring system for the acoustic signature and temperature of conveyor belt idler rollers is provided, such as... Figure 1 As shown, it includes a distributed temperature-measuring fiber optic cable 1, a distributed auscultation fiber optic cable 2, a fiber optic temperature analysis host 6, a fiber optic auscultation analysis host 7, and a monitoring host 8. The two ends of the distributed temperature-measuring fiber optic cable 1 are connected to the fiber optic temperature analysis host 6, forming a closed loop. The fiber optic temperature analysis host 6 determines whether there are any temperature anomalies in the loop based on the signals from the distributed temperature-measuring fiber optic cable 1. The two ends of the distributed auscultation fiber optic cable 2 are connected to the fiber optic auscultation analysis host 7, forming a closed loop. The fiber optic auscultation analysis host 7 determines whether there are any vibration anomalies in the loop based on the signals from the distributed auscultation fiber optic cable 2. The fiber optic auscultation analysis host 7 and the fiber optic temperature analysis host 6 are respectively connected to the monitoring host 8. The monitoring host 8 determines whether a malfunction has occurred in the idler roller operation based on the temperature signals from the fiber optic temperature analysis host 6 and the vibration signals from the fiber optic auscultation analysis host 7.

[0078] In some embodiments, the application of an integrated online monitoring system for the acoustic signature and temperature of conveyor belt idlers is as follows: Figure 2 As shown, the distributed temperature-measuring optical fiber 1 is wound inside the temperature-measuring clamp 4, which is located on the outside of the conveyor roller shaft 3 of the conveyor belt; the distributed auscultation optical fiber 2 is clipped onto the frame of the conveyor belt, i.e., the conveyor belt bracket 5; the distributed temperature-measuring optical fiber 1 is connected to the optical fiber temperature analysis host 6; the distributed auscultation optical fiber 2 is connected to the optical fiber auscultation analysis host 7; and the optical fiber temperature analysis host 6 and the optical fiber auscultation analysis host 7 are respectively connected to the monitoring host 8.

[0079] For a single conveyor belt idler shaft, as an example, such as Figure 3 As shown, the distributed temperature measuring fiber 1 is set inside the temperature measuring fixture 4, which is located at the bolt position or slot position on the outside of the conveyor belt idler shaft 3. The distributed auscultation fiber 2 is close to or in contact with the conveyor belt idler shaft 3.

[0080] This design, through distributed fiber optic sensing technology, accurately monitors the vibration and temperature changes of the idlers, promptly identifies potential faults, and ensures comprehensive monitoring of the entire conveyor belt system, avoiding blind spots in traditional manual inspections. It reduces reliance on manual inspections, lowers manpower input, and improves inspection efficiency. Real-time monitoring and early warning systems promptly detect abnormalities in the idlers, preventing safety accidents and production stoppages caused by malfunctions. This improves the operating efficiency and safety of the conveyor belt, and further promotes the application of intelligent monitoring technology in coal production and transportation.

[0081] In some embodiments, a method for integrated online monitoring of acoustic signature and temperature of conveyor belt idler rollers is as follows: Figure 4 As shown, the process includes the following steps: emitting a first laser beam with pulses, which is transmitted through a distributed stethoscope fiber; receiving scattered light returned from the distributed stethoscope fiber, locating a specific idler area based on Rayleigh scattering, and converting it into a vibration signal; emitting a second laser beam with pulses, which is transmitted through a distributed temperature-sensing fiber; receiving scattered light returned from the distributed temperature-sensing fiber, locating a specific idler area based on Raman scattering, and converting it into a temperature signal; performing data cleaning and fusion processing on the vibration signal and temperature signal to obtain fused features for analysis and training; determining whether a fault has occurred in the idler operation based on the training results; and issuing an alarm signal when a fault occurs in the idler operation.

[0082] In some embodiments, a method for integrated online monitoring of acoustic signature and temperature of conveyor belt idler rollers includes the following steps: a fiber optic auscultation analysis host emits a first laser beam with pulses, and the first laser beam is transmitted through a distributed auscultation fiber; the fiber optic auscultation analysis host receives scattered light returned from the distributed auscultation fiber, locates a specific idler roller area based on Rayleigh scattering, converts it into a vibration signal, and transmits it to the monitoring host; a fiber optic temperature analysis host emits a second laser beam with pulses, and the second laser beam is transmitted in a distributed temperature measurement fiber; the fiber optic temperature analysis host receives scattered light returned from the distributed temperature measurement fiber, locates a specific idler roller area based on Raman scattering, converts it into a temperature signal, and transmits it to the monitoring host; the monitoring host receives the vibration signal and temperature signal, performs data cleaning and fusion processing, obtains fused features for analysis and training, determines whether a fault has occurred in the idler roller operation based on the training results, and issues an alarm signal when a fault occurs in the idler roller operation. As an example, the integrated online monitoring device for the acoustic signature and temperature of conveyor belt idlers is implemented using any embodiment of the integrated online monitoring method for the acoustic signature and temperature of conveyor belt idlers; or, the integrated online monitoring method for the acoustic signature and temperature of conveyor belt idlers is implemented using any embodiment of the integrated online monitoring device for the acoustic signature and temperature of conveyor belt idlers, that is, the integrated online monitoring method for the acoustic signature and temperature of conveyor belt idlers adopts any embodiment of the integrated online monitoring system for the acoustic signature and temperature of conveyor belt idlers. It is understood that, since any embodiment of the integrated online monitoring system for the acoustic signature and temperature of conveyor belt idlers is adopted, the integrated online monitoring method for the acoustic signature and temperature of conveyor belt idlers also has the beneficial technical effects of the integrated online monitoring system for the acoustic signature and temperature of conveyor belt idlers, which will not be elaborated here.

[0083] Before the fiber optic auscultation analysis host emits the first laser beam with pulses, the integrated online monitoring method for acoustic signature temperature of conveyor rollers also includes the step of pre-setting the fiber optic auscultation analysis host: distributed auscultation fibers are set on the frame of the conveyor; similarly, before the fiber optic temperature analysis host emits the second laser beam with pulses, the integrated online monitoring method for acoustic signature temperature of conveyor rollers also includes the step of pre-setting the fiber optic temperature analysis host: distributed temperature measuring fibers are set on the roller shaft of the conveyor.

[0084] To improve the accuracy of fault reporting, in some embodiments, the monitoring host preprocesses and extracts features from vibration and temperature signals to ensure data quality and consistency. As an example, the monitoring host inputs the preprocessed and feature-extracted vibration and temperature signals into a multi-sensor fusion module, which then feeds them into a convolutional neural network for deep analysis training to improve the accuracy of alarm signals.

[0085] This design, on the one hand, allows the monitoring host to preprocess and extract features from vibration and temperature signals, effectively filtering out noise interference and standardizing data formats to ensure the quality and consistency of input data, laying a reliable foundation for subsequent analysis. On the other hand, the processed signals are input into a multi-sensor fusion module, where deep analysis training is performed through convolutional neural networks to fully explore the correlation features between vibration and temperature data, achieving deep fusion and intelligent judgment of multi-source information, significantly improving the accuracy and reliability of fault identification. Furthermore, it relies on deep learning algorithms to autonomously learn fault characteristics, reducing reliance on human experience, lowering the probability of false alarms and missed alarms, and enhancing the system's adaptability to complex working conditions. Finally, intelligent analysis and accurate alarms can shorten fault handling time, reduce operation and maintenance costs, effectively ensure the stable operation of the conveyor belt, and help improve production efficiency and equipment safety.

[0086] To facilitate timely fault handling by users, in some embodiments, the integrated online monitoring method for the acoustic signature and temperature of conveyor belt idlers further includes the step of: when a fault occurs during idler operation, adding the faulty area and fault status of the idler to the alarm signal based on vibration and temperature signals. In some embodiments, when a fault occurs during idler operation, the monitoring host sends an alarm signal to a preset client. In some embodiments, when a fault occurs during idler operation, the monitoring host adds the faulty area and fault status of the idler to the alarm signal based on vibration and temperature signals. In some embodiments, an integrated online monitoring method for the acoustic signature and temperature of conveyor belt idlers is as follows: Figure 5 As shown, the process includes the following steps: emitting a first laser beam with pulses, which is transmitted through a distributed stethoscope fiber; receiving scattered light returned from the distributed stethoscope fiber, locating a specific idler roller area based on Rayleigh scattering, and converting it into a vibration signal; emitting a second laser beam with pulses, which is transmitted through a distributed temperature-sensing fiber; receiving scattered light returned from the distributed temperature-sensing fiber, locating a specific idler roller area based on Raman scattering, and converting it into a temperature signal; performing data cleaning and fusion processing on the vibration and temperature signals to obtain fused features for analysis and training, and determining whether a fault has occurred in the idler roller operation based on the training results; and, in the case of a fault in the idler roller operation, including the faulty area and fault status of the idler roller in the alarm signal based on the vibration and temperature signals. Other embodiments follow the same principle and will not be elaborated further.

[0087] As an example, the monitoring host receives vibration and temperature signals, performs data cleaning and fusion processing, obtains fused features for analysis and training, including the construction of model training datasets and preprocessing methods; specific steps are illustrated below.

[0088] First, the model training dataset is constructed based on actual operating scenarios, such as the real operating scenario of conveyor belt idlers in open-pit coal mines. Multi-source sensor data is collected synchronously through fiber optic distributed vibration sensors, acoustic acquisition modules, and high-precision temperature sensors. The collected scenarios cover the normal operating conditions of the idlers and various typical fault conditions.

[0089] Secondly, vibration and temperature signals are received to acquire data. During this process, the three types of sensor data are aligned with a unified timestamp to ensure that the vibration, acoustic signature, and temperature data at the same sampling time are matched in time sequence, avoiding fusion failure caused by multimodal data misalignment. For the acquired raw time-series data, outlier removal, fixed-duration framing, data normalization, and data augmentation are performed sequentially: abnormal data points caused by strong electromagnetic interference in the mine and sudden mechanical disturbances are removed using mathematical statistics methods; continuous time-series signals are overlapped and framed to expand the sample size; all sensor data are uniformly normalized to the [0,1] interval to eliminate the dimensional differences between different sensor data; and data augmentation is performed through time-domain offset and Gaussian noise superposition to improve the model's scene generalization ability.

[0090] Next, the preprocessed sample dataset is randomly divided into training set, validation set and test set in a ratio of 7:2:1. The dataset is labeled using a method that matches the working condition label with the sample one by one, and a special sample set for training the multi-sensor fusion model is constructed.

[0091] In some embodiments, vibration and temperature signals are cleaned and fused to obtain fused features for analysis and training. This includes the following steps: In some embodiments, vibration and temperature signals are normalized and feature extracted to obtain cleaned signals. These cleaned signals are then fused using a multi-sensor fusion method to obtain fused features. The fused features are then input into a convolutional neural network for analysis and training to obtain training results. In some embodiments, data normalization and feature extraction include the following steps: Preprocessing vibration and temperature signals, dividing them into training and testing sets; extracting features from the vibration and temperature signals in the training set, inputting these features into a training model for training, and using the training results as the cleaned signals; dividing the testing set into a registration set and a validation set, extracting features from the vibration and temperature signals in the registration and validation sets respectively, comparing and verifying them, and using the verification results as the cleaned signals.

[0092] The following example illustrates the process of inputting fused features into a convolutional neural network for analysis and training. In some embodiments, inputting fused features into a convolutional neural network for analysis and training includes the following steps.

[0093] First, the model training environment was set up. The multi-branch one-dimensional convolutional neural network model mentioned above was constructed based on the PyTorch / TensorFlow deep learning framework, and the model structure was initialized. The model training hyperparameters were configured, with the Adam adaptive optimizer used as the optimizer, the initial learning rate set to 1×10, and a dynamic learning rate decay mechanism adopted. The cross-entropy loss function was used for multi-class classification, the batch size was set to 32, and the maximum number of training epochs was set to 100. At the same time, an early stopping mechanism was configured to automatically terminate training when the recognition accuracy on the validation set did not improve for 10 consecutive iterations to prevent the model from overfitting.

[0094] Secondly, load the divided training set, validation set, and test set data, and simultaneously input the three types of paired time-series sample data of fiber optic vibration, acoustic signature, and temperature to ensure that each set of training samples is a spatiotemporally aligned multimodal fusion sample.

[0095] Next, an iterative training process is executed; in each iteration, multimodal sample data is input in batches, and single-modal fault features are extracted layer by layer through three independent feature extraction branches; the three extracted single-modal features are spliced ​​and fused to generate global multi-source fusion features; after feature optimization is completed through a fully connected layer, the idler condition classification prediction results are output through the output layer.

[0096] Then, the cross-entropy loss between the model's prediction results and the actual working condition labels is calculated. The error is then propagated layer by layer through the backpropagation algorithm. Combined with the Adam optimizer, the weight parameters and bias parameters of each layer of the network are iteratively updated to continuously optimize the model's feature extraction and fusion capabilities.

[0097] After each training round, the model's recognition accuracy and loss value are evaluated by calling the validation set data. Training logs are recorded synchronously, and hyperparameters such as the number of convolutional kernels, learning rate, and Dropout coefficients are dynamically fine-tuned based on the validation set results to balance the model's recognition accuracy with the real-time inference speed at the edge.

[0098] Finally, after training is terminated, the model file corresponding to the optimal weight parameters is saved, and the model performance is finally verified through the test set to obtain the trained multi-sensor fusion convolutional neural network model, which can be deployed in the online monitoring system of conveyor belt idlers in open-pit coal mines to realize real-time status monitoring and fault identification.

[0099] In this way, through a multi-branch CNN structure and a dedicated training process, feature-level deep fusion of three types of sensor data—fiber optic vibration, acoustic signature, and temperature—is achieved. This solves the technical defects of traditional single-sensor monitoring, such as poor anti-interference capability, single fault identification dimension, and weak adaptability to complex open-pit coal mine scenarios. It effectively ensures the real-time performance, accuracy, and stability of the idler roller online monitoring system.

[0100] As an example, such as Figure 6 As shown, the integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers includes the following steps: Acquiring sound data through distributed optical fibers, the optical fiber auscultation analysis host uses Rayleigh scattering for positioning, converts the data into vibration signals, and transmits them to the monitoring host; the monitoring host converts the vibration signals, i.e., sound vibration signals, or simply acoustic vibration signals; simultaneously, acquiring temperature data through distributed optical fibers, the optical fiber temperature analysis host uses Raman scattering for positioning, converts the data into temperature signals, and transmits them to the monitoring host; the monitoring host converts the temperature signals; the monitoring host performs data normalization processing, feature extraction, and uses a multi-sensor fusion convolutional neural network, i.e., a multi-feature fusion neural network, to determine whether the idler roller has malfunctioned, and if so, issues a fault alarm, i.e., a fault warning. As an example, the training process of the multi-sensor fusion neural network is as follows: Figure 7 As shown, for the sound vibration and temperature datasets, a training set is formed, features are extracted and the model is trained, and a test set is formed, which is divided into a registration set and a validation set. Features are extracted separately for each set, and the training model is combined to perform multi-sensor fusion convolution. Based on the results of multi-sensor fusion convolution, it is determined whether the idler roller has malfunctioned. If so, a fault warning is issued.

[0101] This design serves several purposes. First, the monitoring host pushes alarm signals to preset clients when idler rollers malfunction, providing real-time reminders to maintenance personnel for timely handling, shortening fault response time, and preventing the escalation of potential hazards. Second, the alarm signals simultaneously carry fault area and status information, enabling precise fault location and status visualization, significantly improving maintenance efficiency. Third, distributed auscultation and temperature measurement optical fibers synchronously collect acoustic vibration and temperature data, which are then located and converted into corresponding signals via Rayleigh scattering and Raman scattering, respectively, achieving synchronous transmission and unified processing of multi-source data. Fourth, after normalizing and extracting features from the data, the monitoring host employs multi-sensor fusion convolutional neural network deep analysis, combined with standardized training processes to optimize the model, effectively improving fault identification accuracy and early warning reliability, reducing false alarms and missed alarms, decreasing reliance on manual inspections, ensuring the safe and continuous operation of the conveyor belt, and improving overall production efficiency and equipment reliability.

[0102] As an example, the integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers includes the following steps.

[0103] The fiber optic auscultation analysis host emits a pulsed laser beam, which is transmitted through distributed auscultation fibers. When the rollers vibrate during operation, the vibrations act on the outer wall of the fiber. According to the Rayleigh scattering principle, this vibration causes a change in the phase of the Rayleigh scattered light inside the fiber. The fiber optic auscultation analysis host receives the scattered light returning from the fiber, calculates the specific location of the signal source based on the signal return time, pinpointing it to a specific roller area, and converts it into a vibration signal that is transmitted to the monitoring host.

[0104] The fiber optic temperature analyzer emits pulsed laser beams that propagate through distributed temperature-sensing optical fibers. As the idler rollers operate, friction generates heat. According to the Raman scattering principle, when this heat is conducted to the optical fibers wound around the roller shaft, the temperature change causes a frequency shift in the scattered light. The fiber optic temperature analyzer receives these scattered light signals, calculates the specific area of ​​the idler roller from which the signal originates, and converts it into a temperature signal, which is then transmitted to the monitoring unit.

[0105] Upon receiving acoustic and temperature signals, the monitoring host first performs preprocessing and feature extraction to ensure data quality and consistency. The cleaned signals are then input into the multi-sensor fusion module, and the fused features are passed to a convolutional neural network for deep analysis training. This allows the system to infer the current fault condition of the idler roller and send the current operating mode to the client.

[0106] When the client receives the idler operation status transmitted from the monitoring host, if a fault signal is detected, it will output the corresponding fault warning and the fault area of ​​the idler, so that maintenance personnel can perform timely and accurate maintenance and repair on the idler.

[0107] This design enables real-time online monitoring of the sound signature and temperature of the conveyor belt rollers, allowing for timely detection of equipment anomalies, reducing the risk of malfunctions, and improving equipment safety and reliability. Furthermore, the analysis of sound signature and temperature data provides crucial information for predictive maintenance, optimizing maintenance strategies and reducing costs. Finally, it enables real-time monitoring and fault diagnosis of conveyor belts, enhancing equipment safety and maintenance efficiency, and has broad application prospects.

[0108] It should be noted that other embodiments of this application also include an integrated online monitoring system and method for the acoustic signature and temperature of conveyor belt idlers, formed by combining the technical features of the above embodiments.

[0109] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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, they should be considered to be within the scope of this specification.

[0110] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the scope of protection of this application. Therefore, the patent protection scope of this application should be determined by the appended claims.

Claims

1. A tape machine roller voiceprint temperature integrated online monitoring system, characterized in that, This includes distributed auscultation fiber optic cables, distributed temperature measurement fiber optic cables, fiber optic auscultation analysis host, fiber optic temperature analysis host, and monitoring host; The distributed auscultation fiber is mounted on the frame of the tape machine, and both ends of the distributed auscultation fiber are connected to the fiber optic auscultation analysis host. The fiber optic auscultation analysis host is configured to emit a first laser beam with pulses into the distributed auscultation fiber and receive scattered light returned from the distributed auscultation fiber, locate a specific roller area based on Rayleigh scattering, and convert it into a vibration signal; The distributed temperature-measuring optical fiber is mounted on the idler roller shaft of the conveyor belt, and both ends of the distributed temperature-measuring optical fiber are connected to the optical fiber temperature analysis host. The fiber optic temperature analysis host is configured to emit a second laser beam with pulses into the distributed temperature measurement fiber and receive scattered light returned from the distributed temperature measurement fiber, locate a specific idler roller area based on Raman scattering, and convert it into a temperature signal; The fiber optic auscultation analysis host and the fiber optic temperature analysis host are respectively connected to the monitoring host. The monitoring host determines that the idler roller has malfunctioned based on the vibration signal and temperature signal, and issues an alarm signal.

2. The belt conveyor idler acoustic signature temperature integrated online monitoring system of claim 1, wherein, The distributed auscultation fiber is configured to be clipped onto the frame of the tape machine.

3. The integrated online monitoring system for acoustic signature and temperature of conveyor belt idler rollers according to claim 1, characterized in that, The distributed temperature-measuring optical fiber is installed at the bolt position or slot position on the outside of the idler roller shaft of the conveyor belt.

4. The integrated online monitoring system for acoustic signature and temperature of conveyor belt idler rollers according to claim 1, characterized in that, The integrated online monitoring system for acoustic signature and temperature of the conveyor belt roller also includes a clamp, in which the distributed temperature measuring optical fiber is wound. The clamp is configured to magnetically attach to the outside of the conveyor belt roller shaft.

5. The integrated online monitoring system for acoustic signature and temperature of conveyor belt idler rollers according to claim 4, characterized in that, The clamp is configured to magnetically attach to a bolt position or a slot position on the outside of the idler shaft of the conveyor belt.

6. The integrated online monitoring system for acoustic signature and temperature of conveyor belt idler rollers according to any one of claims 1 to 5, characterized in that, The fiber optic auscultation analysis host and the fiber optic temperature analysis host are integrated into one unit.

7. A method for integrated online monitoring of acoustic signature and temperature of conveyor belt idler rollers, characterized in that, Including the following steps: A first laser beam with pulses is emitted, and the first laser beam is transmitted through distributed stethoscope fibers; It receives scattered light returned from the distributed auscultation fiber, locates specific idler roller areas based on Rayleigh scattering, and converts it into vibration signals; A second laser beam with pulses is emitted, and the second laser beam is transmitted in a distributed temperature-sensing optical fiber; It receives scattered light returned from the distributed temperature-measuring optical fiber, locates specific idler roller areas based on Raman scattering, and converts it into a temperature signal; The vibration and temperature signals are cleaned and fused to obtain fused features for analysis and training. Based on the training results, it is determined whether the idler roller has malfunctioned, and an alarm signal is issued when the idler roller malfunctions.

8. The integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers according to claim 7, characterized in that, The integrated online monitoring method for sound and temperature of conveyor belt idler rollers also includes the following steps: when the idler roller malfunctions, the fault area and fault status of the idler roller are included in the alarm signal based on the vibration signal and temperature signal.

9. The integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers according to claim 7, characterized in that, The vibration and temperature signals are cleaned and fused to obtain fused features for analysis and training. The steps include: normalizing the vibration and temperature signals and extracting features to obtain cleaned signals; fusing the cleaned signals using a multi-sensor fusion method to obtain fused features; and inputting the fused features into a convolutional neural network for analysis and training to obtain training results.

10. The integrated online monitoring method for acoustic signature and temperature of conveyor belt idler rollers according to claim 9, characterized in that, Data normalization and feature extraction include the following steps: The vibration and temperature signals were preprocessed and divided into training and testing sets. Feature extraction is performed on vibration and temperature signals in the training set, and the results are used as the cleaned signals. The test set is divided into a registration set and a validation set. Vibration signals and temperature signals in the registration set and validation set are extracted and compared for verification. The verification results are used as the cleaned signals.

Images