Complex Multimodal Phase Noise Suppression Method for Phase-Sensitive Optical Time Domain Reflectometer

Abstract

The present invention belongs to the field of distributed optical fiber acoustic wave sensing technology, and is a method for suppressing complex multimodal phase noise of a phase-sensitive optical time domain reflectometer, including: obtaining a noise-containing signal by subjecting two beat frequency electrical signals output by a balanced photodetector to I / Q demodulation and unwrapping processing; performing SVMD decomposition to obtain the decomposed modes, and judging the correlation between the modes and the noise-containing signal through the SWPCC method to retain the components with high correlation at the window positions corresponding to the modes; traversing all modes, and accumulating all the components to be retained according to the corresponding window positions to obtain a phase noise cancellation result. In the process of calculating the correlation between each mode and the original signal in the present invention, the influence of irrelevant components in the mode on the correlation coefficient is eliminated, and the signal-to-noise ratio of the DAS signal is significantly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of distributed fiber optic acoustic sensing, and particularly relates to a method for suppressing complex multimodal phase noise of a phase-sensitive optical time domain reflectometer. Background Art

[0002] Distributed fiber optic sensing (DAS) realizes long-distance, long-time, and large-scale sensing with the characteristics of low cost, corrosion resistance, and anti-interference, and is widely used in various fields such as oil and gas exploration, pipeline monitoring, traffic detection, disaster prediction, and geophysical applications. The phase-sensitive optical time domain reflectometer ( -OTDR) system is one of the key research directions in this field. This system injects highly coherent laser light into the fiber under test, and the vibration of the medium under test is reflected by the change of Rayleigh backscattered light (RBS). These vibration signals are often affected by various noises. In the DAS system, common noises include light source frequency drift (LFD), beat noise, electrical noise, and fading noise. In addition, in practical applications, the environmental noise is often at a relatively high level. In order to achieve more accurate sensing, it is necessary to suppress the noise in the signal to improve the signal-to-noise ratio.

[0003] In order to suppress the noise in the signal and improve the signal-to-noise ratio of the DAS signal, only signals with simple frequency components are considered in the existing methods. When the frequency components of the signal become complex, the results will deteriorate seriously. The reason is that the effective components of the signal may be randomly and irregularly distributed in all modes, and the amount of information contained in each mode is unknown. This leads to the situation that for a certain mode, it may have only a very short segment of useful signal, resulting in the calculated correlation coefficient being lower than the threshold and being discarded, causing the loss of signal components. Summary of the Invention

[0004] The present invention provides a method for suppressing complex multimodal phase noise of a phase-sensitive optical time domain reflectometer, which solves the problem of signal component loss and is difficult to be applied to complex engineering applications, improves the signal-to-noise ratio of the signal, and is beneficial to accurate distributed acoustic sensing applications.

[0005] The present invention is implemented as follows. A method for suppressing complex multimodal phase noise of a phase-sensitive optical time domain reflectometer includes the following steps:

[0006] S1: The two beat electrical signals output by the balanced photodetector are subjected to I / Q demodulation and unwrapping processing to obtain a noise-containing signal , where is time;

[0007] S2: The noise-containing signal is subjected to SVMD decomposition to obtain the decomposed modes , where represents the a mode indicating the total number of modes;

[0008] S3: Determine the mode and the correlation with the noisy signal to retain the components with high correlation at the corresponding window positions; Traverse all modes and accumulate all the components to be retained according to the corresponding window positions to obtain the phase noise cancellation result .

[0009] Furthermore, specifically included in S1 are:

[0010] The two beat electrical signals that balance the output of the photodetector are expressed as:

[0011] (1),

[0012] wherein, is the optoelectronic conversion coefficient of the balanced photodetector; and represent two orthogonal components of the local oscillator light, is the component of the local oscillator light, is the component of the local oscillator light, and represent two orthogonal components of the signal light, is the component of the signal light, is the component of the signal light; is the angular frequency, where is the modulation frequency, is the sampling frequency, is the phase of the signal light, is the phase of the local oscillator light, represents the component of the beat electrical signal, represents the component of the beat electrical signal;

[0013] Perform I / Q quadrature demodulation on the two beat electrical signals to obtain 2 orthogonal components in the polarization direction , and 2 orthogonal components in the polarization direction , expressed as: , ,

[0014] (2),​

[0015] Among them, represents the local oscillator light, represents the signal light. The orthogonal components in two polarization directions are superimposed to obtain the in-phase signal component and the quadrature in-phase signal component , is the orthogonal projection angle: ,

[0016] According to the in-phase signal component and the quadrature in-phase signal component the winding phase is obtained:

[0017] (4),

[0018] The unwinding process is as follows:

[0019] (5),

[0020] (6),

[0021] Among them, is the difference between the original phase and the winding phase at time , represents the difference between the original phase and the winding phase at time is the unwound phase at time i.e., the signal with noise, represents the winding phase at time represents the winding phase at time is the unwinding threshold.

[0022] Furthermore, the process of SVMD decomposition in S2 includes: dividing the signal with noise into two parts: (7)

[0023] Among them, represents the residual signal. Dividing the residual signal into two parts gives:

[0024] (8),

[0025] Among them, represents the sum of the previous -1 modes, Indicates the unprocessed part in the frequency variable of the Fourier transform. For the noisy signal perform Fourier transform to obtain the unwrapped phase in the frequency domain , mode transform to obtain the mode in the frequency domain . According to Parseval's theorem, the sum of the frequency-domain energies of the decomposed parts is equal to the energy of the original signal. For , let be the sampling point, indicates the th sampling point's mode , at time is expressed as , indicates the th sampling point's mode , is the center frequency of the mode, indicates the center frequency of the mode of indicates the center frequency of the mode of indicates the center frequency of the mode; indicates the th mode's Lagrange multiplier, indicates the th sampling point's Lagrange multiplier, and let , , , respectively, iteratively calculate , and based on equations (9), (10), and (11):

[0026] (9),

[0027] (10),

[0028] (11),

[0029] where is the result after Fourier transform of ; is the result after Fourier transform of ; is the balancing factor; is the th mode's Lagrange multiplier of the sampling point; denote the Fourier transform result at the nth sampling point, be the modal Lagrange multiplier of the sampling point, be the update coefficient. Finally, calculate the mode according to the inverse Fourier transform .

[0030] Furthermore, S3 specifically includes:

[0031] Let the starting position of the window be , the width of the sliding window be , the PCC threshold be , calculate the PCC value of the mode from the position to and the noisy signal , denote the arithmetic mean of the mode in this position range as , the arithmetic mean of the noisy signal in this position range as , , denote the signal of the noisy signal at the nth sampling point in the window, then the calculation formula of the PCC value is as follows:

[0032] ,

[0033] Compare the PCC value with the set PCC threshold . If , it indicates that the mode at the current window position should be discarded; if , it indicates that the mode at the current window position is retained, that is:

[0034] ,

[0035] where is the phase denoising result; denote the denoising result at the window position, that is the denoising result from the sampling point to the sampling point; denote the th mode at the window position, that is the modal component from the sampling point to the sampling point.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0037] The method proposed by the present invention solves the problem that the existing methods are difficult to be applied to complex engineering applications, which enables the "modal decomposition - correlation coefficient" denoising strategy to be applied to more DAS applications with complex signal components. Traditional methods such as EMD - PCC and VMD - PCC are effective for simple signals. When the frequency components of the signal become complex, a certain mode after decomposition does not necessarily contain all the effective information. In a certain mode, valuable signal components often only appear in several small segments. At this time, calculating the correlation based on all the information of a certain mode for judgment will obviously cause errors. In this regard, the method of the present invention optimizes the process of calculating the correlation between each mode and the original signal, eliminates the influence of irrelevant components in the mode on the correlation coefficient, and significantly improves the signal - to - noise ratio of DAS signals.

[0038] The present invention uses successive variational mode decomposition (SVMD) as the mode decomposition method, which solves the problem that it is difficult to determine the number of modes by the VMD method when the signal components are unknown. On the other hand, by introducing a Pearson correlation coefficient (PCC) calculation method with a moving window, the correlation coefficient calculation strategy is improved. The local correlation between each mode and the corresponding position of the original signal is judged segment by segment, and the highly correlated local components are accurately retained without being affected by the irrelevant components in the mode, realizing accurate phase denoising. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is the overall flowchart of the method provided by the embodiment of the present invention;

[0040] Figure 2 is the structural schematic diagram of the phase - sensitive optical time - domain reflectometer provided by the embodiment of the present invention;

[0041] Figure 3 is the 5Hz noisy phase signal provided by the embodiment of the present invention;

[0042] Figure 4 is the denoising result of the 5Hz signal provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0044] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0045] As Figure 1 shown, a complex multimodal phase noise suppression method for a phase-sensitive optical time domain reflectometer includes the following steps:

[0046] S1: Subject the two beat electrical signals output by a balanced photodetector (DAS system) to I / Q demodulation and unwrapping processing to obtain a noise-containing signal , where

[0047] is time; S2: Perform SVMD decomposition on the noise-containing signal to obtain the decomposed modes , where represents the th mode, and

[0048] represents the total number of modes; SVMD refers to successive variational mode decomposition; S3: Judge the correlation between the mode and the noise-containing signal by the SWPCC method to retain the components with high correlation at the window positions corresponding to the mode ; traverse all modes and accumulate all the components to be retained according to the corresponding window positions to obtain the phase noise cancellation result ; SWPCC refers to the Pearson correlation coefficient processing method.

[0049] See Figure 2 shown, the phase-sensitive optical time domain reflectometer based on the present invention is composed of multiple components, Figure 2The solid lines in it are the optical paths, and the dashed lines are the circuits, including a narrow-linewidth laser 1, an optical fiber coupler 2, an acousto-optic modulator 3, an erbium-doped fiber amplifier 4, an optical fiber circulator 5, a polarization diversity module 6, a balanced photodetector one 7, a balanced photodetector two 8, a data acquisition module 9, a data demodulation module 10, a signal generator 11, a measured optical fiber one 12, a measured optical fiber two 13, and a piezoelectric transducer 14. The signal generator 11 is connected to the piezoelectric transducer 14. In the phase-sensitive optical time-domain reflectometer, the central frequency of the narrow-linewidth laser 1 is 1550 nm, and the emitted laser is divided into two parts by a 90:10 optical fiber coupler 2, namely 90% signal light and 10% local oscillator light. The signal light is modulated into pulsed light by the acousto-optic modulator 3 and then amplified by the front-end erbium-doped fiber amplifier 4. The amplified pulsed light enters the measured optical fiber one 12 through the optical fiber circulator 5, and a piezoelectric transducer 14 is connected between the measured optical fiber one 12 and the measured optical fiber two 13. The signal light undergoes backscattering in the measured optical fiber one 12, the measured optical fiber two 13, and the piezoelectric transducer 14 to generate Rayleigh scattering signal light. These signal lights pass through the optical fiber circulator 5 and are input into the polarization diversity module 6 together with the local oscillator light to form beat signals in two polarization directions. These beat signals are converted into electrical signals by the balanced photodetector one 7 and the balanced photodetector two 8 and finally enter the data acquisition module 9 and the data demodulation module 10;

[0050] In this embodiment, the repetition frequency of the narrow-linewidth laser 1 is set to 2000 Hz. The lengths of the measured optical fiber one 13 and the measured optical fiber two 14 are 1 km. The piezoelectric transducer 12 is arranged between the measured optical fiber one 12 and the measured optical fiber two 13, and its driving voltage is set to a 5 Hz sine signal with Gaussian noise of 2Vpp, which is used to generate a 5 Hz vibration signal on the optical fiber;

[0051] Among them, the method of the present invention is applied to a processor that can receive the output of the balanced photodetector, including but not limited to mobile phones, laptop computers, desktop computers, etc. A method for suppressing complex multimodal phase noise of a phase-sensitive optical time-domain reflectometer provided in this embodiment. Specifically, in S1, the beat electrical signal output by the balanced photodetector is expressed as:

[0052] (1),

[0053] Among them, is the optoelectronic conversion coefficient of the balanced photodetector; and represent two orthogonal components of the local oscillator light, is the component of the local oscillator light, is the component of the local oscillator light, and Represent two orthogonal components of the signal light, is the component of the signal light, is the component of the signal light; is the angular frequency, where is the modulation frequency, is the sampling frequency, is the phase of the signal light, is the phase of the local oscillator light, represents the component of the beat frequency electrical signal, represents the component of the beat frequency electrical signal;

[0054] Perform I / Q orthogonal demodulation on the two beat frequency electrical signals to obtain 2 orthogonal components in the polarization direction , and orthogonal components in the polarization direction ; , , expressed as:

[0055] (2),

[0056] where represents the local oscillator light, represents the signal light. Superimpose the orthogonal components in the two polarization directions to obtain the in-phase signal component and the quadrature-phase signal component , is the orthogonal projection angle:

[0057] (3),

[0058] According to the in-phase signal component and the quadrature-phase signal component obtain the unwrapped phase :

[0059] (4),

[0060] The unwrapping process is as follows:

[0061] (5),

[0062] (6),

[0063] where is the difference between the original phase and the unwrapped phase at time, represents The difference between the original phase of time and the winding phase, is the unwound phase of time, i.e., the noisy signal, ; denotes the winding phase of time, denotes the winding phase of time, is the unwinding threshold.

[0064] As Figure 3 shown. It can be obtained from Figure 3 that, except for the 5 Hz sine signal at the position in, there is a large amount of random noise.

[0065] In S2: The process of SVMD decomposition includes: dividing the noisy signal into two parts:

[0066] (7),

[0067] wherein, denotes the residual signal, and dividing the residual signal into two parts gives:

[0068] (8),

[0069] wherein, denotes the sum of the previous -1 modes, denotes the unprocessed part in, denotes the frequency variable of the Fourier transform. Performing the Fourier transform on the noisy signal gives the noisy signal in the frequency domain , and mode is transformed to mode in the frequency domain. According to Parseval's theorem, the sum of the frequency-domain energies of the decomposed parts is equal to the energy of the original signal. For , let be the sampling point, denotes the th sampling point of mode , when is denoted as , denotes the th sampling point of mode , is the central frequency of the mode, denotes the central frequency of the mode of, represents the central frequency of the mode of represents the central frequency of the mode of ; represents the Lagrange multiplier of the mode, represents the Lagrange multiplier at the -th sampling point, and let , , , and are iteratively calculated respectively based on formulas (9), (10), and (11):

[0070] (9),

[0071] (10),

[0072] (11),

[0073] where is the result after Fourier transform of ; is the result after Fourier transform of ; is the balance factor; is the Lagrange multiplier of the -th mode at the -th sampling point after variable substitution; represents the Fourier transform result of at the -th sampling point, is the Lagrange multiplier of the -th mode at the -th sampling point after variable substitution,

[0074] In S3: The purpose of the SWPCC method is to judge the correlation between and the noisy signal to retain the components with high correlation. Let the starting position of the window be , the width of the sliding window be , the PCC (Pearson correlation coefficient) threshold be , calculate the PCC value of the mode between the position from to the position of the noisy signal , denote the mode The arithmetic mean within this position range is , including the noise signal The arithmetic mean within this position range is , If it represents the signal at the nth sampling point of the noise-containing signal within the window, then the calculation formula for the PCC value is as follows:

[0075] ,

[0076] Compare the PCC value with the set PCC threshold . If , it indicates that the mode at the current window position should be discarded; if , it indicates that the mode at the current window position is retained, that is:

[0077] ,

[0078] Among them, is the result of phase noise cancellation; represents the noise cancellation result at the window position, that is the noise cancellation result from the sampling point to the sampling point; represents the th mode at the window position, that is the modal component from the sampling point to the sampling point.

[0079] As Figure 4 shown, it is the result after noise cancellation. By comparing Figure 3 and Figure 4 , it can be concluded that after the noise cancellation process of the present invention, the noise in the noise-containing signal is effectively eliminated.

[0080] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A complex multimodal phase noise suppression method for a phase-sensitive optical time domain reflectometer, characterized in that, Including the following steps: S1: Demodulate the two beat electrical signals output by the balanced photodetector through I / Q demodulation and de-winding processing to obtain a signal with noise , is time; S2: For the noisy signal perform SVMD decomposition to obtain the decomposed modes , where represents the th mode and represents the total number of modes; S3: Determine the mode through the SWPCC method The correlation with the noisy signal to retain the components with high correlation at the corresponding window positions of the mode; traverse all modes and accumulate all the components to be retained according to the corresponding window positions to obtain the phase denoising result ; ; Specifically included in S3: Set the starting position of the window as , the width of the sliding window is , the PCC threshold is , calculate the PCC value of the mode from the position to within the range and the noisy signal The formula for the PCC value is as follows: Denote the arithmetic mean of the mode within this position range as , and the arithmetic mean of the noisy signal within this position range as . Denote the signal at the nth sampling point of the noisy signal within the window, then the formula for the PCC value is as follows:​ ， Compare the PCC value with the set PCC threshold . If , it indicates that the modality at the current window position should be discarded; if , it indicates that the modality at the current window position is retained, i.e.: ， Among them, is the phase noise cancellation result; represents the noise cancellation result at the window position, that is, the noise cancellation result from the sampling point to the sampling point; represents the th mode at the window position, that is, the modal component from the sampling point to the sampling point.

2. The complex multi-modal phase noise suppression method for a phase-sensitive optical time domain reflectometer according to claim 1, wherein Specifically included in S1: The two beat electrical signals output by the balanced photodetector are expressed as: （1）， Among them, is the optoelectronic conversion coefficient of the balanced photodetector; and represent two orthogonal components of the local oscillator light, is the component of the local oscillator light, is the component of the local oscillator light, and represent two orthogonal components of the signal light, is the component of the signal light, is the component of the signal light; is the angular frequency, where is the modulation frequency, is the sampling frequency, is the phase of the signal light, is the phase of the local oscillator light, represents the component of the beat frequency electrical signal, represents the component of the beat frequency electrical signal; I / Q quadrature demodulation is performed on two beat frequency electrical signals to obtain two orthogonal components in the polarization direction 、 and the orthogonal components in the polarization direction 、 , expressed as: （2）， Among them, represents the local oscillator light, represents the signal light, and the orthogonal components in two polarization directions are superimposed to obtain the in-phase signal component and the quadrature-phase signal component , is the orthogonal projection angle: ， Based on the in-phase signal component and the quadrature-phase signal component the wrapped phase is obtained as follows: （4）， The unwrapping process is as follows: （5）， （6）， Wherein, is the difference between the original phase and the winding phase of time, denotes the difference between the original phase and the winding phase of time, is the phase after unwinding of time, i.e., the signal with noise, ; denotes the winding phase of time, denotes the winding phase of time, is the unwinding threshold.

3. A complex multi-modal phase noise suppression method for a phase-sensitive optical time domain reflectometer according to claim 2, characterized in that The process of SVM decomposition in S2 includes: dividing the noisy signal into two parts: （7）， Among them, represents a residual signal, and the residual signal is obtained by dividing it into two parts: （8）， Among them, represents the sum of the previous -1 modes, represents the unprocessed part in represents the frequency variable of the Fourier transform. The unwrapped phase in the frequency domain is obtained by performing the Fourier transform on the noisy signal After that, the mode is transformed to obtain the mode in the frequency domain . According to Parseval's theorem, the sum of the energy in the frequency domain of each decomposed part is equal to the energy of the original signal. For , let be the sampling points, represents the mode at the th sampling point . When it is expressed as . represents the mode at the th sampling point is the center frequency of the mode, represents the center frequency of the mode of represents the center frequency of the mode of represents the center frequency of the mode of ; represents the Lagrange multiplier of the th mode, represents the Lagrange multiplier at the th sampling point, and let , , , respectively, based on formulas (9), (10), and (11), iteratively calculate , and : （9）， （10）， （11）， Among them, is the result after Fourier transform; is the result after Fourier transform; is the balance factor; is the Lagrange multiplier of the modal sampling point after variable substitution; represents the Fourier transform result at the nth sampling point, is the Lagrange multiplier of the modal sampling point after variable substitution, is the update coefficient. Finally, the mode is calculated according to the inverse Fourier transform.

Images