Roller abnormality diagnosis system based on LFM-DAS

Abstract

The application belongs to the technical field of optical fiber distributed sensing, and discloses a kind of based on LFM-DAS's roll abnormal diagnosis system, wherein, narrow linewidth laser output laser is divided into signal light and reference light two beams after first coupler, signal light is modulated as LFM optical pulse after LFM pulse modulation module, and enters sensing optical fiber after optical circulator;Rayleigh backscattering light generated in sensing optical fiber is incident to the first input end of second coupler after output by optical circulator, and sensing optical fiber is fixedly arranged on belt conveyor roll;Reference light is incident to the second input end of second coupler after polarization controller;The output end of second coupler is connected with signal acquisition processing module;Signal acquisition processing module is used to collect beat frequency signal of reference light and Rayleigh backscattering light output in sensing optical fiber, and whether roll abnormality occurs and abnormal type is judged according to the beat frequency signal collected, the present application can realize the real-time detection of roll small abnormal event, and the calculation is small, and the speed is fast.

Description

Technical Field

[0001] This invention belongs to the field of fiber optic distributed sensing technology, specifically relating to an idler roller anomaly diagnosis system based on LFM-DAS (Linear Frequency Modulated Pulse-Fiber Distributed Acoustic Sensing). It is used to solve the problems of insufficient anti-interference capability, poor real-time performance and severe fading phenomenon in traditional DAS solutions when monitoring the operating status of belt conveyors in complex coal mine environments, thereby realizing real-time monitoring and early warning of belt conveyor idler roller eccentricity and shaft breakage faults. Background Technology

[0002] Belt conveyors, with their advantages of low maintenance costs, large conveying capacity, and reliable operation, have become a key continuous transportation equipment in coal mining. Idler rollers are one of the key structural components of belt conveyors, responsible for supporting and rolling the conveyor belt. Due to long-term use and environmental factors, idler rollers are prone to failure. If not detected in time, the belt conveyor may be forced to shut down, and in severe cases, this can lead to fires and safety accidents. Therefore, monitoring and diagnosing idler rollers is crucial. Traditional idler roller inspection usually relies on manual methods, subjectively inspecting the operating status of belt conveyor idler rollers through visual inspection and listening. This primitive inspection method is labor-intensive, has a time lag, and cannot detect changes while the belt conveyor is in operation. In contrast, fiber optic sensing technology has advantages such as strong resistance to electromagnetic interference, strong corrosion resistance, easy installation, and long-distance transmission. Furthermore, optical fibers can convert minute vibrations on the surface of an object into changes in propagation path and light intensity, providing excellent sensing capabilities for vibration and stress changes. This makes it highly suitable for complex environments with long-distance belt conveyors and a large number of idler rollers.

[0003] In related technologies, Chinese patent document CN112173636A discloses a method for detecting belt conveyor idler roller faults using an inspection robot. This method intelligently identifies and analyzes frame deformation images acquired through prior video recording and inputs them into an alarm predictor. Information fusion reduces the false alarm rate of anomaly diagnosis and lowers resource consumption during on-site belt conveyor inspections. However, this method requires installing a mobile camera on the belt conveyor, which can lead to problems such as unclear images, significant image changes, and missing information during high-speed belt conveyor movement. Chinese patent document CN119262732A discloses a belt conveyor idler roller anomaly detection system based on fiber optic auscultation. This method converts the collected audio data from the idler roller's operation into vibration data, and then performs time-domain and frequency-domain analysis in the data processing unit. This enables fault warning by analyzing the sound signals of the belt conveyor idler roller. However, this method requires converting the sound signal to the vibration signal, which may result in errors during the conversion process, and the signal conversion is very time-consuming, resulting in poor real-time performance.

[0004] Traditional DAS systems suffer from fading, leading to detection blind spots. Therefore, in order to achieve intelligent operation and maintenance of belt conveyors and comprehensive perception of idler status, and to improve the continuity, real-time performance, and accuracy of idler fault diagnosis, it is necessary to provide a belt conveyor idler abnormal operation status monitoring method based on DAS system. This method can directly extract and process vibration information and overcome the detection blind spots caused by fading in traditional DAS systems. Summary of the Invention

[0005] To address the problems of high resource consumption, strong lag, monitoring blind spots, and poor real-time performance in existing belt conveyor operation status detection systems based on fiber optic sensing technology, this invention proposes an idler roller anomaly diagnosis system based on LFM-DAS.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: an LFM-DAS-based roller abnormality diagnosis system, including a narrow linewidth laser, a first coupler, an LFM pulse modulation module, a polarization controller, an optical circulator, a second coupler, a sensing optical fiber, and a signal acquisition and processing module.

[0007] The narrow-linewidth laser output is split into two beams, a signal beam and a reference beam, after passing through the first coupler. The signal beam is modulated into an LFM pulse by an LFM pulse modulation module and then enters the sensing fiber through an optical circulator. The Rayleigh backscattered light generated in the sensing fiber is output through the optical circulator and then incident on the first input end of the second coupler. The sensing fiber is fixedly mounted on the conveyor belt roller. The reference beam is incident on the second input end of the second coupler after passing through a polarization controller.

[0008] The output of the second coupler is connected to the signal acquisition and processing module; the signal acquisition and processing module is used to acquire the beat frequency signal of the Rayleigh backscattered light output from the reference light and the sensing fiber, and to determine whether an idler roller abnormality has occurred and the type of abnormality based on the acquired beat frequency signal.

[0009] The specific method by which the signal acquisition and processing module determines whether an idler roller abnormality has occurred and the type of abnormality based on the beat frequency signal is as follows:

[0010] Step 1: Acquire beat frequency signals and preprocess the beat frequency data;

[0011] Step 2: Fill the preprocessed data along the distance axis of the sensing fiber and the vibration period time axis to obtain a two-dimensional time domain matrix;

[0012] Step 3: Perform a Hilbert transform on the two-dimensional time-domain matrix to obtain the two-dimensional envelope matrix;

[0013] Step 4: Determine if there is a timing offset in the signal of the two-dimensional envelope matrix. If there is no timing offset, it is determined that no roller abnormality has occurred. If a sinusoidal impact-type timing offset occurs, determine whether the peak-to-peak value corresponding to the timing offset exceeds the first threshold. If it exceeds the threshold, it is determined to be a broken shaft condition in roller abnormality. If it does not exceed the threshold, it is determined that no roller abnormality has occurred. If a linear timing offset occurs, perform a short-time Fourier transform on the two-dimensional envelope matrix point by point and window by window to convert it to the frequency domain to obtain the time spectrum. Determine whether the peak frequency in the time spectrum exceeds the second threshold. If it exceeds the threshold, it is determined to be an eccentric condition in roller abnormality. If it does not exceed the threshold, it is determined that no roller abnormality has occurred.

[0014] In step 1, the specific method for data preprocessing is as follows:

[0015] Perform analog-to-digital conversion on the beat frequency signal;

[0016] The gradient signal is obtained by filtering the beat frequency signal after analog-to-digital conversion. The filtering formula is as follows:

[0017] ;

[0018] in, and They represent and The beat frequency signal acquired at each time step; δ represents the gradient threshold, and G(τ) represents the time step. The gradient signal;

[0019] Automatic gain compensation is performed on the filtered gradient signal. The compensation formula is as follows:

[0020] ;

[0021] in, This indicates the compensated signal. The gain coefficient within window k is determined by the average power of the gradient signal within window k.

[0022] In step 1, the gain coefficient The calculation formula is:

[0023] ;

[0024] in, This represents the average power of the gradient signal within window k. Indicates the target power.

[0025] The specific method by which the signal acquisition and processing module determines whether a roller malfunction has occurred and the type of malfunction based on the beat frequency signal is as follows:

[0026] Step 1: Acquire beat frequency signals and preprocess the beat frequency data;

[0027] Step 2: Fill the preprocessed data along the distance axis of the sensing fiber and the vibration period time axis to obtain a two-dimensional time domain matrix;

[0028] Step 3: Perform a Hilbert transform on the two-dimensional time-domain matrix to obtain the envelope matrix;

[0029] Step 4: Perform a point-by-point, window-by-window short-time Fourier transform on the envelope matrix to convert it to the frequency domain to obtain the time spectrum. Then, perform phase demodulation and phase unwinding to obtain the phase signal. Determine whether an idler roller abnormality has occurred and the type of abnormality based on the phase signal.

[0030] The LFM pulse modulation module includes an LFM signal generator, a Mach-Zehnder modulator, a bias controller, and a beam splitter. The LFM signal generator generates an LFM signal to drive the Mach-Zehnder modulator to modulate the signal light output from the first coupler into an LFM optical pulse. The LFM optical pulse is split into two beams by the beam splitter. One beam is biased to zero by the intensity modulator of the Mach-Zehnder modulator and to the quadrature point by the phase shifter, and the other beam is output to the optical circulator.

[0031] The signal acquisition and processing module includes a balanced photodetector, a data acquisition unit, and a data processor.

[0032] The balanced photodetector is used to receive the beat frequency signal of the reference light and the Rayleigh backscattered light output from the sensing fiber, and after photoelectric conversion, it is sent to the data acquisition unit. After data acquisition by the data acquisition unit, it is sent to the data processor.

[0033] The data processor is used to determine whether an idler roller malfunction has occurred and the type of malfunction based on the beat frequency signal.

[0034] The data acquisition unit is a high-speed oscilloscope, and the sensing fiber is a single-mode fiber.

[0035] The LFM-DAS-based roller abnormality diagnosis system further includes a first erbium-doped fiber amplifier and a second erbium-doped fiber amplifier.

[0036] The first erbium-doped fiber amplifier is located at the output end of the LFM pulse modulation module, and the second erbium-doped fiber amplifier is located at the output end of the optical circulator, used to amplify the Rayleigh backscattered light signal generated in the sensing fiber.

[0037] The first coupler is a bias-maintaining coupler.

[0038] Compared with the prior art, the present invention has the following advantages:

[0039] (1) This invention uses linear frequency modulated pulses as the detection signal of a distributed fiber optic acoustic sensing system. By utilizing the linear change of the pulse signal frequency over time, interference fading noise is eliminated, thereby improving the accuracy of signal monitoring. Furthermore, by performing one-dimensional gradient filtering on the acquired two-dimensional DAS data and applying automatic gain control to enhance the weak signals in the data, the signal-to-noise ratio can be further improved, enabling real-time detection of minor abnormal events on the idler roller.

[0040] (2) The present invention adopts the time shift characteristics unique to the LFM-DAS system and combines them with the physical model corresponding to the belt conveyor roller fault to characterize the roller abnormal event-fiber coupling mechanism. It can characterize the vibration signal by the time shift signal in the collected beat frequency signal, and then analyze the real-time working status of the roller in real time. It does not require phase data demodulation, the corresponding calculation is small, and the analysis speed is fast.

[0041] In summary, this invention proposes an idler roller anomaly diagnosis system based on LFM-DAS. By analyzing the correspondence between idler roller anomaly events and the scattering timing and time shift characteristics of linear frequency modulated pulses, the influence of interference fading on the system is eliminated, thereby achieving real-time detection of minor idler roller anomalies. The analysis results are accurate and the diagnosis speed is fast. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the structure of an LFM-DAS-based roller abnormality diagnosis system provided in an embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of the physical model of abnormal events of belt conveyor rollers in an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the time-shift characteristics generated by an LFM-DAS-based roller anomaly diagnosis system in an embodiment of the present invention.

[0045] Figure 4 This is a flowchart of roller abnormality diagnosis in an embodiment of the present invention;

[0046] Figure 5 This is a time-series waterfall plot corresponding to the two-dimensional envelope matrix obtained during the simulation of the ramp signal in this embodiment of the invention;

[0047] Figure 6 This is a time-series waterfall plot corresponding to the two-dimensional envelope matrix obtained during the simulation of the sinusoidal impact signal in this embodiment of the invention;

[0048] In the figure: 1 Narrow linewidth laser, 2 LFM pulse modulation module, 3 Sensing fiber, 4 Signal acquisition and processing module, 5 First coupler, 6 Polarization controller, 7 First erbium-doped fiber amplifier, 8 Optical circulator, 9 Second erbium-doped fiber amplifier, 10 Second coupler;

[0049] 203 - Arbitrary waveform generator; 204 - Mach-Zehnder modulator; 205 - Bias controller; 206 - Beam splitter; 310 - Drive signal generator; 311 - Power amplifier; 313 - Piezoelectric ceramic tube; 416 - Balanced photodetector; 417 - High-speed real-time oscilloscope; 418 - Data processor. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] like Figure 1 As shown, this embodiment of the invention provides an LFM-DAS-based roller abnormality diagnosis system, including a narrow linewidth laser 1, a first coupler 5, an LFM pulse modulation module 2, a polarization controller 6, an optical circulator 8, a second coupler 10, a sensing fiber 3, and a signal acquisition and processing module 4.

[0052] The laser output from the narrow linewidth laser 1 is split into two beams, a signal beam and a reference beam, after passing through the first coupler 5. The signal beam is modulated into an LFM light pulse by the LFM pulse modulation module 2 and then enters the sensing fiber 3 through the optical circulator 8. The Rayleigh backscattered light generated in the sensing fiber 3 is output from the optical circulator 8 and then incident on the first input end of the second coupler 10. The sensing fiber 3 is fixedly mounted on the conveyor belt roller. The reference beam is incident on the second input end of the second coupler 10 after passing through the polarization controller 6.

[0053] The output of the second coupler 10 is connected to the signal acquisition and processing module 4; the signal acquisition and processing module 4 is used to acquire the beat frequency signal of the Rayleigh backscattered light output from the reference light and the sensing fiber, and to determine whether an abnormality of the idler roller has occurred and the type of abnormality based on the acquired beat frequency signal.

[0054] Specifically, in this embodiment, the signal acquisition and processing module 4 includes a balanced photodetector 416, a data acquisition unit 417, and a data processor 418. The balanced photodetector 416 is used to receive the beat frequency signal of the reference light and the Rayleigh backscattered light output from the sensing fiber, and after photoelectric conversion, it is sent to the data acquisition unit 417. After data acquisition by the data acquisition unit 417, the data is sent to the data processor 418. The data processor 418 is used to determine whether an idler roller abnormality has occurred and the type of abnormality based on the beat frequency signal.

[0055] Specifically, in this embodiment, the data acquisition device 417 is a high-speed oscilloscope, and the sensing fiber 3 is a single-mode fiber.

[0056] In this embodiment, the polarization controller 6 is used to adjust the polarization of the reference light output by the first coupler 5 so that it matches the polarization of the Rayleigh scattering probe light output by the circulator 8. The reference light after passing through the polarization controller 6 is incident on the other input end of the second coupler 10.

[0057] In addition, in this embodiment, a piezoelectric ceramic tube 313 is set on the sensing optical fiber 3 to simulate the abnormal impact of the idler roller. The driving signal emitted by the driving signal generator 310 is amplified by the power amplifier 311 and drives the piezoelectric ceramic tube 313 to vibrate, simulating the abnormal impact of the idler roller on the sensing optical fiber 3.

[0058] like Figure 2 As shown, in this embodiment, the drive signal generator 310 outputs a ramp excitation signal to simulate the eccentricity of the idler roller. The essence of the idler roller eccentricity fault is the uneven mass distribution of the rotating body, leading to periodic changes in centrifugal force. Its vibration characteristics are manifested as a linearly increasing-decreasing inertial force related to the rotational speed. To accurately simulate this physical process, a triangular wave or sawtooth wave voltage driven by a frequency synchronized with the idler roller rotational speed is used to drive the piezoelectric ceramic tube 313. The signal amplitude linearly rises to a peak value over time and then drops sharply (or gradually decreases), corresponding to the dynamic process where the centrifugal force is greatest when the eccentric mass point rotates to its lowest position and suddenly decreases when it rotates to its highest position. The rising edge of the ramp wave simulates the linear accumulation of centrifugal force when the eccentric mass rotates from the high point to the low point, and the falling edge simulates the release of force after the mass crosses the lowest point. Signal frequency f 0 must satisfy f0 = n / 6 ( n (This refers to the roller rotation speed). By applying a corresponding signal to the piezoelectric ceramic tube 313, a periodic sawtooth displacement waveform with the same frequency as the ramp signal can be measured. The spectrum shows the fundamental frequency and its harmonic components, which is consistent with the actual eccentricity fault characteristics.

[0059] like Figure 2As shown, in this embodiment, the drive signal generator 310 outputs a sinusoidal impact excitation signal to simulate the roller shaft breakage condition. The instant of shaft breakage generates a transient impact force, accompanied by free decaying vibration of the remaining shaft segment. Its dynamic response can be modeled as a damped oscillation system. The excitation signal is driven by a single-cycle or short-time-range sinusoidal signal. The main lobe of the sinusoidal wave simulates the stress wave release at the instant of shaft breakage. During the impact phase, the main lobe of the sinusoidal wave simulates the stress wave release at the instant of shaft breakage; during the decay phase, the side lobe envelope simulates the vibration decay at the fracture point due to frictional damping.

[0060] like Figure 3 As shown in this embodiment, when a strain changes at a certain position in the sensing fiber 3, the optical path difference will change. Since the frequency change of the linear frequency modulated pulse can compensate for the change in optical path difference caused by the strain, a time shift occurs. Therefore, the Rayleigh scattering traces of different detection pulses completely overlap in the non-disturbance event region, while a time shift phenomenon occurs in the strain region. Figure 3 The solid green line and the dashed orange line show a time shift in the disturbance region.

[0061] Specifically, in this embodiment, as Figure 4 As shown, the method by which the signal acquisition and processing module 4 determines whether an idler roller abnormality has occurred and the type of abnormality based on the beat frequency signal is as follows:

[0062] Step 1: Acquire beat frequency signals and preprocess the data.

[0063] In step 1, the specific method for data preprocessing is as follows:

[0064] Step 1.1 Perform analog-to-digital conversion on the beat frequency signal;

[0065] Step 1.2 Filter the beat frequency signal after analog-to-digital conversion to obtain the gradient signal. The filtering formula is:

[0066] (1)

[0067] in, and They represent and The beat frequency signal acquired at each time step; δ represents the gradient threshold, and G(τ) represents the time step. The gradient signal;

[0068] Step 1.3 performs automatic gain compensation on the filtered gradient signal. The compensation formula is as follows:

[0069] (2)

[0070] in, This indicates the compensated signal. The gain coefficient within window k is determined by the average power of the gradient signal within window k.

[0071] Specifically, gain coefficient The calculation formula is:

[0072] (3)

[0073] in, This represents the average power of the gradient signal G(τ) within window k. This represents the target power. Specifically, the formula for calculating the average power is:

[0074] (4)

[0075] M is the number of window points, which can be calculated based on the oscilloscope's sampling rate.

[0076] Step 2: Fill the preprocessed data into a two-dimensional matrix along the distance axis of the sensing fiber and the vibration period time axis to obtain a two-dimensional time domain matrix.

[0077] The preprocessed data signal is a one-dimensional signal with time on the horizontal axis and intensity on the vertical axis. By dividing the time axis into multiple segments through the transmission interval of two adjacent LFM pulses, the signal of each LFM pulse across the entire sensing fiber can be obtained. Therefore, the preprocessed one-dimensional signal can be reconstructed into a two-dimensional time-domain matrix M including distance d and time t. Here, distance d is defined by the propagation time of the light pulse, reflecting the spatial location of the vibration event and is key to locating roller anomalies. The time axis t, in units of vibration period, reflects the characteristics of vibration variation over time at a specific location.

[0078] Step 3: Perform a Hilbert transform on the two-dimensional time-domain matrix M to obtain the two-dimensional envelope matrix.

[0079] Using the two-dimensional time-domain matrix M obtained in step 2 as the processing object, perform Hilbert transform on it to eliminate phase interference in the original time-domain signal, thereby obtaining an envelope signal that can intuitively reflect the change in vibration intensity.

[0080] The specific calculation process is as follows: For the time series signal M(d, t) corresponding to any distance d in the two-dimensional time domain matrix M, its analytic signal H[M(d, t)] is obtained through Hilbert transform, and its expression is:

[0081] H [M (d, t)]=M (d, t)+j・N(d, t); (5)

[0082] Where j is the imaginary unit, N(d, t) represents the Hilbert transform signal of the time series M(d, t), M(d, t) and N(d, t) are orthogonal in the time domain, and:

[0083] N(d, t)=IFFT {H(d, t)·FFT [M (d, t)]}; (6)

[0084] Where H(d, t) is the Hilbert transform operator, IFFT and FFT represent the inverse Fourier transform and Fourier transform, respectively. Then, the modulus of the analytic signal is taken to obtain the vibration envelope signal corresponding to the distance d.

[0085] E (d, t) = |H [M (d, t)]|; (7)

[0086] Here, E(d, t) represents the vibration envelope signal corresponding to distance d. By integrating all vibration envelope signals corresponding to distances, a two-dimensional envelope matrix E of distance d and time t can be formed.

[0087] Step 4: Determine whether there is a timing offset in the signal in the two-dimensional envelope matrix. If there is no timing offset, it means that there is no strain or temperature change and no roller abnormality has occurred.

[0088] (1) If a sinusoidal impulsive timing offset phenomenon occurs in the signal of the two-dimensional envelope matrix, it is necessary to determine whether the peak-to-peak value of the timing offset exceeds the set first threshold. If it exceeds, it is determined to be a broken shaft condition in the idler roller abnormality; if it does not exceed, it may be a slight deformation of the idler roller caused by cargo loading, and in this case, it is determined that no idler roller abnormality has occurred. The waterfall diagram of the sinusoidal impulsive timing offset phenomenon is as follows: Figure 6 As shown, its temporal offset exhibits periodicity. The first threshold is a distance threshold.

[0089] (2) If a linear time-series offset occurs in the signal within the two-dimensional envelope matrix, the envelope matrix needs to be transformed into the frequency domain using a point-by-point, window-by-window short-time Fourier transform to obtain the time spectrum. It is then determined whether the peak frequency in the time spectrum exceeds the second threshold. If it does, it is identified as an eccentric condition in the idler roller abnormality; if it does not exceed the threshold, it may be a minor jamming of the idler roller during cargo pulling, in which case it is determined that no idler roller abnormality has occurred. The second threshold is a frequency threshold.

[0090] Furthermore, in this embodiment, the location of the abnormal idler roller can be accurately located based on the distance d and time t corresponding to the time offset.

[0091] In this embodiment, when the idler roller deviates or breaks, a time shift occurs in the disturbance event region of the two-dimensional envelope matrix. Therefore, the presence or absence of time shift can be used to preliminarily determine whether a fault has occurred. For sinusoidal impact-type time shift phenomena, whether the peak-to-peak value corresponding to the time shift exceeds a first threshold can be used to directly determine whether an idler roller shaft breakage has occurred. For linear time shift phenomena, it may be due to minor jamming of the idler roller during cargo pulling. Therefore, time-frequency conversion is required, and whether the frequency peak value in the time-frequency spectrum exceeds a second threshold is used to determine whether the idler roller is eccentric. In other words, this invention utilizes the characteristic that external disturbances (temperature changes or strain) cause time shifts in the Rayleigh scattering signal when using LFM pulses as probe light. It directly judges idler roller anomalies through time shift without needing phase demodulation of the two-dimensional envelope matrix, which can greatly improve the judgment speed and reduce the amount of data processing.

[0092] In addition, in this embodiment, the phase signal corresponding to the vibration signal can be extracted by performing phase demodulation on the time spectrum obtained by short-time Fourier transform. The waveform of the fiber strain caused by the roller abnormality (slope wave signal / sinusoidal shock wave signal) over time can be recovered from the phase signal, and finally the abnormality of the roller and the type of abnormality can be determined. However, this method has a relatively large computational load and long demodulation time. Specifically, the time spectrum is obtained by short-time Fourier transform, and the phase angle φ_d(t) corresponding to the LFM modulation frequency component is obtained from the time spectrum. The phase angle φ_d(t) reflects the phase modulation information caused by vibration at time t (corresponding to position d). φ_d(t) is the wrapped phase, and its value is restricted to the main value range of (-π, π] or [0, 2π). Since the real phase change caused by vibration is usually continuous and may exceed 2π radians, and φ_d(t) is the wrapped phase, a 2π jump will occur when the real phase crosses the ±π or 2π boundary, that is, phase entanglement. To recover the true, continuous phase change Φ_d(t), it is necessary to unwrap the phase angle φ_d(t) of the wrapped phase. After unwrapping, the phase signal generated by the roller vibration measured by the DAS system based on linear frequency modulation pulses can be extracted by phase difference operation before and after the disturbance region, thereby recovering the strain change caused by roller anomalies (slope wave signal / sinusoidal shock wave signal).

[0093] Furthermore, such as Figure 1As shown, in this embodiment, the LFM pulse modulation module 2 includes an LFM signal generator 203, a Mach-Zehnder modulator 204, a bias controller 205, and a beam splitter 206. The LFM signal generator 203 generates an LFM signal to drive the Mach-Zehnder modulator 204 to modulate the signal light output from the first coupler 5 into an LFM optical pulse. The LFM optical pulse is split into two beams by the beam splitter 206. One beam is biased to the zero point by the intensity modulator of the Mach-Zehnder modulator 204 and the phase shifter is biased to the quadrature point by the control signal output by the bias controller 205. The other beam is output to the optical circulator 8.

[0094] The beam splitter 206 is a 90:10 optical coupler, with 1% of its portion connected to the bias controller 205 to provide feedback and 99% of its portion outputting LFM optical pulses.

[0095] Specifically, in this embodiment, the LFM signal generator is used to generate a digital LFM single-pulse signal with a pulse width of 100ns and a repetition period of 220μs. The LFM signal generator 203 can be an arbitrary waveform generator. The narrow linewidth laser 1 is used to output laser light with a center wavelength of 1550.12 nm and a linewidth of less than 0.1kHz.

[0096] Furthermore, the LFM-DAS-based roller abnormality diagnosis system of this embodiment also includes a first erbium-doped fiber amplifier 7 and a second erbium-doped fiber amplifier 9; the first erbium-doped fiber amplifier 7 is disposed at the output end of the LFM pulse modulation module 2, and the second erbium-doped fiber amplifier 9 is disposed at the output end of the optical circulator 8, for amplifying the Rayleigh backscattered light signal generated in the sensing fiber 3.

[0097] Specifically, the first erbium-doped fiber amplifier 107 is a pulse-type erbium-doped fiber amplifier, and the second erbium-doped fiber amplifier 9 is a small-signal erbium-doped fiber amplifier.

[0098] Specifically, the first coupler 5 is a 1×2 polarization-maintaining coupler with a beam splitting ratio of 90:10, where 90% of the beam is used as signal light and 10% is used as reference light.

[0099] Specifically, the second coupler 10 is a 1×2 optical coupler with a splitting ratio of 50:50.

[0100] The data processing flow and principle of this embodiment are described below.

[0101] In this embodiment, the LFM optical pulse output by the LFM pulse modulation module is a rectangular pulse representing intensity, and its expression is:

[0102] (7)

[0103] in, Indicates LFM optical pulse signal, The amplitude of the LFM optical pulse is represented by T, and the duration of the LFM optical pulse is represented by T. This represents the instantaneous frequency of the LFM optical pulse, and , Let represent the initial optical frequency of the LFM optical pulse, t represent time, and μ represent the chirp rate of the LFM optical pulse, satisfying the following relationship: , This indicates the cutoff frequency of the LFM optical pulse.

[0104] LFM optical pulses are injected into sensing fiber 3 to generate Rayleigh backscattered light. This light is output from optical circulator 8, amplified by a second erbium-doped fiber amplifier 9, and then coupled to the reference light output from polarization controller 6 in second coupler 10. The optical signal output from second coupler 115 is detected by balanced photodetector 416. Any scattering point within sensing fiber 3 will generate a scattering trace, and the output light intensity detected by balanced photodetector 416 is also considered. The reference light signal... The expression is:

[0105] (8)

[0106] in, Indicates the amplitude of the reference light. This indicates the phase at the initial moment.

[0107] Rayleigh backscattered signal output from optical circulator 8 The expression is:

[0108] (9)

[0109] Where r is the Rayleigh scattering coefficient, a is the attenuation coefficient for propagation within the optical fiber, c represents the speed of light in vacuum, n represents the refractive index, and T... z Let φ be the transmission time from the laser source to position z in the optical fiber and then to the receiver of the detection signal. z This represents the phase information at position z, and rect represents a rectangular function.

[0110] The expression for the beat frequency interference light signal obtained by the balanced photodetector 416 is:

[0111] (10)

[0112] Where E(t) = E LO(t) +E s(t)This represents the mixed optical signal of the reference light and Rayleigh backscattered light obtained through coupler 10. In this embodiment, a linear frequency modulated pulse is introduced into the DAS system. The frequency change of the pulse compensates for the phase change caused by external vibration, resulting in a shift in the time-series signal waterfall plot corresponding to the two-dimensional envelope matrix, which is the time shift phenomenon of LFM-DAS. The phase difference between the scattered signals at any two positions in the sensing fiber 3 is:

[0113] (11)

[0114] in, and These represent the frequencies of the i-th and j-th discrete scattering units, respectively. and These represent the times when the pulse returns to the signal receiver after passing through the i-th and j-th discrete scattering units, respectively. From the above equations, the sensing principle can be described from three aspects: time perturbation and optical path difference, optical path difference and coherent Rayleigh scattering, and coherent Rayleigh scattering and time-domain shift. Specifically, perturbation of the measured event signal affects the propagation path or medium of light, causing a change in the optical path difference, which in turn alters the coherence of the coherent Rayleigh scattering signal, ultimately causing a time-domain shift in the coherent Rayleigh scattering signal. Therefore, the relationship between strain ε and time shift Δt can be obtained:

[0115] (12)

[0116] From the above equation, it can be seen that as long as a certain position of the sensing fiber 3 is subjected to a certain dynamic strain disturbance, it will cause a change in strain, denoted by Δε, and thus generate a change in the coherent time domain signal. The horizontal offset is such that the magnitude of the vibration can be determined by the time shift offset.

[0117] Therefore, this invention measures the beat frequency signal of the Rayleigh backscattered signal in the sensing fiber, and based on the time shift characteristics of the beat frequency signal, it can deduce the characteristics of the external strain disturbance, thereby identifying whether there is an abnormality in the idler roller and the type of abnormality.

[0118] The vibration of the piezoelectric ceramic tube 313 is driven by a ramp signal to simulate an idler roller eccentricity fault. The time-series waterfall plot corresponding to the obtained two-dimensional envelope matrix is ​​shown below. Figure 5 As shown, the amplitude of the driving signal gradually increases from (a) to (c). Figure 5 This indicates that eccentric conditions occurring at different times within the monitoring period can be observed using time-series waterfall plots. It can be seen that when the idler roller is eccentric, there is a significant shift in the corresponding time-series waterfall plot. Specifically, the shift in (a) occurs between 0 and 10 ms, the shift in (b) occurs between 40 and 50 ms, and the shift in (c) occurs between 10 and 20 ms.

[0119] The time-series waterfall plot corresponding to the two-dimensional envelope matrix obtained by simulating a roller shaft breakage fault through vibration of a piezoelectric ceramic tube 313 driven by a sinusoidal wave signal is shown below. Figure 6 As shown, the amplitude of the driving signal gradually decreases from (a) to (c). It can be seen that the peak-to-peak value corresponding to the time shift also gradually decreases during the process of the amplitude decreasing (h1>h2>h3), which proves that different degrees of shaft breakage can be observed through the time-frequency diagram.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A roller anomaly diagnosis system based on LFM-DAS, characterized in that, Includes a narrow linewidth laser (1), a first coupler (5), an LFM pulse modulation module (2), a polarization controller (6), an optical circulator (8), a second coupler (10), a sensing fiber (3), and a signal acquisition and processing module (4); The narrow linewidth laser (1) outputs laser light, which is split into two beams, signal light and reference light, after passing through the first coupler (5). The signal light is modulated into an LFM light pulse by the LFM pulse modulation module (2) and then enters the sensing fiber (3) through the optical circulator (8). The Rayleigh backscattered light generated in the sensing fiber (3) is output through the optical circulator (8) and then incident on the first input end of the second coupler (10). The sensing fiber (3) is fixedly mounted on the conveyor belt roller. The reference light is incident on the second input end of the second coupler (10) after passing through the polarization controller (6). The output of the second coupler (10) is connected to the signal acquisition and processing module (4); the signal acquisition and processing module (4) is used to acquire the beat frequency signal of the Rayleigh backscattered light output from the reference light and the sensing fiber, and to determine whether an idler roller abnormality has occurred and the type of abnormality based on the acquired beat frequency signal; the specific method by which the signal acquisition and processing module (4) determines whether an idler roller abnormality has occurred and the type of abnormality based on the beat frequency signal is as follows: Step 1: Acquire beat frequency signals and preprocess the beat frequency data; Step 2: Fill the preprocessed data along the distance axis of the sensing fiber and the vibration period time axis to obtain a two-dimensional time domain matrix; Step 3: Perform a Hilbert transform on the two-dimensional time-domain matrix to obtain the two-dimensional envelope matrix; Step 4: Determine whether there is a timing offset in the signal in the two-dimensional envelope matrix. If there is no timing offset, it is determined that no roller abnormality has occurred. If a sinusoidal impact-type timing offset occurs, determine whether the peak-to-peak value corresponding to the timing offset exceeds the first threshold. If it does, it is determined to be a broken shaft condition in the roller abnormality. If the limit is not exceeded, it is determined that no roller abnormality has occurred; If a linear time-series offset occurs, the two-dimensional envelope matrix is ​​transformed into the frequency domain by performing a short-time Fourier transform point by point and window by window to obtain the time spectrum. It is then determined whether the peak frequency in the time spectrum exceeds the second threshold. If it does, it is determined to be an eccentric condition in the roller abnormality; if it does not exceed the threshold, it is determined that no roller abnormality has occurred.

2. The LFM-DAS-based roller abnormality diagnosis system according to claim 1, characterized in that, In step 1, the specific method for data preprocessing is as follows: Perform analog-to-digital conversion on the beat frequency signal; The gradient signal is obtained by filtering the beat frequency signal after analog-to-digital conversion. The filtering formula is as follows: ； in, and They represent and The beat frequency signal acquired at each time step; δ represents the gradient threshold, and G(τ) represents the time step. The gradient signal; Automatic gain compensation is performed on the filtered gradient signal. The compensation formula is as follows: ； in, This indicates the compensated signal. The gain coefficient within window k is determined by the average power of the gradient signal within window k.

3. The LFM-DAS-based roller abnormality diagnosis system according to claim 2, characterized in that, In step 1, the gain coefficient The calculation formula is: ； in, This represents the average power of the gradient signal within window k. Indicates the target power.

4. The LFM-DAS-based roller abnormality diagnosis system according to claim 1, characterized in that, The LFM pulse modulation module (2) includes an LFM signal generator (203), a Mach-Zehnder modulator (204), a bias controller (205), and a beam splitter (206). The LFM signal generator (203) generates an LFM signal to drive the Mach-Zehnder modulator (204) to modulate the signal light output from the first coupler (5) into an LFM optical pulse. The LFM optical pulse is split into two beams by the beam splitter (206). One beam is biased to zero by the intensity modulator of the Mach-Zehnder modulator (204) and the phase shifter is biased to the quadrature point by the control signal output by the bias controller (205). The other beam is output to the optical circulator (8).

5. The LFM-DAS-based roller abnormality diagnosis system according to claim 1, characterized in that, The signal acquisition and processing module (4) includes a balanced photodetector (416), a data acquisition unit (417), and a data processor (418). The balanced photodetector (416) is used to receive the beat frequency signal of the reference light and the Rayleigh backscattered light output from the sensing fiber and perform photoelectric conversion before sending it to the data acquisition unit (417). After data acquisition by the data acquisition unit (417), the data is sent to the data processor (418). The data processor (418) is used to determine whether an abnormality has occurred in the idler roller and the type of abnormality based on the beat frequency signal.

6. The LFM-DAS-based roller abnormality diagnosis system according to claim 1, characterized in that, The data acquisition unit (417) is a high-speed oscilloscope, and the sensing fiber (3) is a single-mode fiber.

7. The LFM-DAS-based roller abnormality diagnosis system according to claim 1, characterized in that, It also includes a first erbium-doped fiber amplifier (7) and a second erbium-doped fiber amplifier (9). The first erbium-doped fiber amplifier (7) is located at the output end of the LFM pulse modulation module (2), and the second erbium-doped fiber amplifier (9) is located at the output end of the optical circulator (8) to amplify the Rayleigh backscattered light signal generated in the sensing fiber (3).

8. The LFM-DAS-based roller abnormality diagnosis system according to claim 1, characterized in that, The first coupler (5) is a bias-maintaining coupler.

Images