Low-temperature crosstalk distributed acoustic sensing system based on antiresonant hollow fiber

Abstract

This application relates to the field of fiber optic sensing technology and discloses a low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow-core fiber. The system includes an optical signal module, a feature extraction and demodulation module, an optical domain calibration module, a noise reduction module, and a compensation and reconstruction module. The sensing medium uses anti-resonant hollow-core sensing fiber to eliminate the thermo-optical effect of solid materials. The system extracts the scattering intensity variance sequence of the interference signal and the original spatial phase. The optical domain calibration module separates the attenuation coefficient along the path and generates a weight mask. The noise reduction module maps the variance sequence to a Kalman filter model and adaptively updates the measurement noise covariance matrix to suppress local burst noise. The compensation and reconstruction module combines the attenuation coefficient and the mask to calibrate the spatial baseline drift, and finally unwraps and reconstructs the acoustic vibration signal. This invention solves the problems of low-frequency temperature crosstalk masking weak strain, difficulty in suppressing local random abrupt noise, and measurement errors caused by spatial baseline drift.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing technology, specifically to a low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow fiber. Background Technology

[0002] Distributed acoustic sensing technology uses optical fiber as the sensing medium. By demodulating the phase change of the backscattered Rayleigh light inside the optical fiber, it can realize continuous spatial positioning and continuous measurement of external strain and vibration signals along the line. Existing distributed acoustic sensing systems generally use solid-core quartz single-mode optical fiber as the sensing link.

[0003] In practical applications such as monitoring minute leaks in long-distance pipelines or monitoring geological deformation, the target strain signals to be extracted are usually concentrated in the low-frequency range. The core material of solid-core quartz single-mode optical fiber has inherent thermal expansion and thermo-optic effects. When the external ambient temperature fluctuates, the equivalent refractive index of the solid medium changes with temperature. This phase drift caused by temperature crosstalk is particularly obvious in the low-frequency range, and the resulting low-frequency phase shift often masks the phase changes caused by weak low-frequency strain.

[0004] To eliminate interference caused by low-frequency temperature drift, existing conventional processing methods mostly rely on back-end digital signal processing, that is, directly applying a fixed high-pass filter algorithm to forcibly filter out low-frequency signal components. While suppressing temperature baseline drift, this processing logic also simultaneously cuts off the effective frequency band of the target low-frequency strain signal, directly limiting the system's ability to monitor low-frequency slow-change phenomena.

[0005] Furthermore, when the local spatial location of an optical fiber is compressed or disturbed, it can cause amplitude fading of the backscattered signal, which manifests as non-uniformly distributed random abrupt noise in the demodulated spatial phase. Existing system denoising algorithms mostly employ conventional filtering models with fixed noise covariance parameters, performing equal-weight smoothing on all nodes in the spatial domain. Such algorithm models lack feedback and linkage mechanisms with the underlying scattering characteristics, and cannot adaptively adjust parameters to address random fluctuations in local light intensity at spatial nodes. This leads to demodulation distortion or subsequent phase unwrapping errors when dealing with local abrupt phase distortions. Simultaneously, the inherent optical power attenuation and local nonlinear disturbances in long-distance sensing links cause static baseline shifts in the spatial dimension. Existing technologies lack spatial calibration and weighted masking methods that incorporate optical fiber-to-optic measurement parameters, resulting in errors in the final reconstructed acoustic vibration signal values. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber. This system solves the problems in existing distributed acoustic sensing systems, such as low-frequency temperature crosstalk severely masking weak strain characteristics, the inability of conventional fixed-parameter filtering algorithms to adaptively suppress local random abrupt noise, and the lack of spatial baseline calibration leading to errors in signal reconstruction.

[0007] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber, comprising: The optical signal module is used to inject the probe light pulse into the anti-resonant hollow core sensing fiber, receive the returned back Rayleigh scattered light and heterodyne mix it with the local oscillator reference light to output an interference electrical signal. The feature extraction and demodulation module is used to extract the backscattered light intensity envelope from the interference electrical signal, calculate the time variance to generate a scattering intensity variance sequence, and demodulate the original spatial phase signal from the interference electrical signal. The optical domain calibration module is used to separate the attenuation coefficient and backscattering coefficient of the anti-resonant hollow-core sensing fiber, and to generate a weighted mask sequence by combining the backscattering coefficient and the scattering intensity variance sequence. The noise reduction module is used to reduce the noise of the original spatial phase signal using Kalman filtering, map and update the measurement noise covariance matrix using the scattering intensity variance sequence, and output the filtered phase signal. The compensation and reconstruction module is used to establish a weighted theoretical baseline drift term using the attenuation coefficient and the weighted mask sequence, subtract the weighted theoretical baseline drift term from the filtered phase signal to obtain a baseline compensation sequence, and unwrap the baseline compensation sequence to reconstruct the acoustic vibration signal.

[0008] Preferably, the optical signal module includes: The emission excitation unit is used to generate continuous light and modulate the continuous light into the probe light pulse, amplify the probe light pulse, and inject it into the anti-resonant hollow core sensing fiber. A sensing medium unit is used to conduct the probe light pulse, which undergoes axial geometric deformation when modulated by external strain and returns the back Rayleigh scattered light carrying axial geometric deformation information. A receiving mixer unit is used to receive the backscattered Rayleigh light and amplify the signal, perform heterodyne mixing operation between the amplified backscattered Rayleigh light and the local oscillator reference light, and output the interference electrical signal.

[0009] Preferably, the feature extraction and demodulation module includes: An envelope extraction unit is used to receive the interference electrical signal and separate the interference electrical signal to obtain the backscattered light intensity envelope corresponding to the spatial distribution node of the anti-resonant hollow sensing fiber. The variance calculation unit is used to obtain the time mean of the backscattered light intensity envelope within a time window, calculate the squared difference between the backscattered light intensity envelope and the time mean, and combine the values ​​to generate the scattering intensity variance sequence. The spatial phase demodulation unit is used to perform a phase demodulation mathematical algorithm on the interference electrical signal, extract and output the original spatial phase signal.

[0010] Preferably, the spatial phase demodulation unit is used to receive the interference electrical signal, separate the interference electrical signal into three voltage signals with a phase difference of 120 degrees between them, perform mathematical calculations on the three voltage signals with a phase difference of 120 degrees between them according to the arctangent demodulation equation, and output the original spatial phase signal.

[0011] Preferably, the optical domain calibration module includes: The parameter measurement and separation unit is used to transmit test pulses to the anti-resonant hollow-core sensing fiber, record the optical power distribution curve along the return path, and obtain the attenuation coefficient and the backscattering coefficient by calculating the optical power distribution curve. An abnormal node identification unit is used to perform data mapping and alignment between the backscattering coefficient and the scattering intensity variance sequence to obtain the numerical status of the backscattering coefficient and scattering intensity variance sequence of the spatial node. The mask generation unit is used to assign weight factors to the spatial nodes according to the numerical state, and to concatenate the weight factors of all the spatial nodes to form the weight mask sequence.

[0012] Preferably, the abnormal node identification unit is used to receive a backscattering anomaly threshold and an intensity variance threshold, perform a comparison operation between the backscattering coefficient and the backscattering anomaly threshold, and perform a comparison operation between the value of the corresponding spatial node in the scattering intensity variance sequence and the intensity variance threshold. When the backscattering coefficient is greater than the backscattering anomaly threshold or the value of the corresponding spatial node is greater than the intensity variance threshold, the corresponding spatial node is marked as a nonlinear interference node.

[0013] Preferably, the noise reduction module includes: The state parameter mapping unit is used to receive the scattering intensity variance sequence and input the scattering intensity variance sequence into the state space model mapping function to generate covariance adjustment parameters. The covariance update unit is used to substitute the covariance adjustment parameter into the Kalman filter and replace the row and column values ​​of the measurement noise covariance matrix node by node in the spatial dimension. The filtering execution unit is used to read the replaced measurement noise covariance matrix, calculate the Kalman gain, substitute the Kalman gain into the filtering equation to perform smoothing processing on the original spatial phase signal, and output the filtered phase signal.

[0014] Preferably, the state parameter mapping unit is used to extract the spatial node values ​​in the scattering intensity variance sequence, multiply the spatial node values ​​with the measurement noise reference parameter to obtain the numerical amplification factor, and output the numerical amplification factor as the covariance adjustment parameter.

[0015] Preferably, the compensation and reconstruction module includes: A drift modeling unit is used to calculate the spatial phase offset using the attenuation coefficient, multiply the weighted mask sequence with the spatial phase offset, and output the weighted theoretical baseline drift term. The difference compensation unit is used to receive the filtered phase signal and the weighted theoretical baseline drift term, perform a subtraction operation on the filtered phase signal and the weighted theoretical baseline drift term, and output the baseline compensation sequence. The untangling and reconstruction unit is used to perform a phase unwinding algorithm on the baseline compensation sequence to remove the tangled phase, and then combine the pure geometric phase stress inverse mapping model to convert the baseline compensation sequence with the tangled phase removed into the acoustic vibration signal.

[0016] Preferably, the unwrapping and reconstruction unit is used to perform a branch cutting algorithm on the baseline compensation sequence, expand the baseline compensation sequence wrapped in the range of negative to positive pi values ​​to obtain continuous phase data, substitute the continuous phase data into the pure geometric phase stress inverse mapping model, reverse solve and output the amplitude and frequency parameters of the acoustic vibration signal.

[0017] This invention provides a low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber. It offers the following advantages: 1. This invention uses an anti-resonant hollow-core sensing fiber sealed at both ends as the sensing medium, so that the optical phase change caused by external strain is determined only by the change in the axial geometric length of the fiber, and the equivalent refractive index of the internal gas remains constant. This eliminates the thermo-elastic-optical effect caused by temperature fluctuations affecting the refractive index in traditional solid-core optical fibers, thereby reducing the interference of temperature crosstalk on low-frequency acoustic signals and ensuring the signal objectivity of the phase demodulation process.

[0018] 2. In the demodulation process, the present invention simultaneously extracts the backscattered light intensity envelope and calculates the scattering intensity variance sequence, and maps it to update the measurement noise covariance matrix in the Kalman filter. This processing uses the data of the local perturbation state of the optical fiber as a priori condition input to the noise reduction algorithm, and dynamically adjusts the Kalman gain point by point at the spatial nodes. Thus, when dealing with the weak backscattering characteristics of hollow media, it can suppress the phase noise of local random abrupt changes by reducing the gain ratio of the nodes.

[0019] 3. This invention utilizes the backscattering coefficient and scattering intensity variance sequence obtained separately to generate a weighted mask sequence, and combines it with the spatially distributed attenuation coefficient to establish a weighted theoretical baseline drift term to compensate for the difference in the filtered phase signal. Furthermore, by allocating weight factors, it actively weakens the nonlinear phase interference caused by non-axial abnormal compression nodes, and simultaneously calculates and deducts the baseline offset accumulated by long-distance links, ensuring the continuity of the phase data output to the subsequent unwrapping and reconstruction modules. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system architecture of the present invention; Figure 2 This is a comparison diagram of the phase response of different sensing media of the present invention under low-frequency temperature interference; Figure 3 This is a comparison diagram of the phase waveforms before and after processing by the noise reduction module of this invention. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and 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.

[0022] Reference Figure 1 The present invention provides a low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber. The system includes: an optical signal module, a feature extraction and demodulation module, an optical domain calibration module, a noise reduction module, and a compensation and reconstruction module.

[0023] The optical signal module serves as the system's underlying hardware and sensing link. It generates probe light pulses and inputs them into the anti-resonant hollow-core sensing fiber. When the anti-resonant hollow-core sensing fiber receives external strain, it deforms and generates backscattered Rayleigh light. The optical signal module receives the backscattered Rayleigh light returning along the fiber and performs heterodyne mixing with the local reference light to output an interference electrical signal carrying modulation information.

[0024] The feature extraction and demodulation module is connected to the optical signal module via a circuit or communication bus to receive the output interference electrical signal. The feature extraction and demodulation module performs parallel dual-branch data processing. In the first branch, it extracts the backscattered light intensity envelope from the interference electrical signal and calculates the time variance based on the time window to generate a spatially distributed scattering intensity variance sequence. In the second branch, it performs a mathematical demodulation algorithm on the interference electrical signal to calculate and output the original spatial phase signal.

[0025] The optical domain calibration module is connected to the feature extraction and demodulation module and the link where the anti-resonant hollow core sensing fiber is located. The optical domain calibration module independently runs the optical time domain reflectance measurement operation, and separates the attenuation coefficient and backscattering coefficient of the anti-resonant hollow core sensing fiber in the spatial domain. The optical domain calibration module receives the scattering intensity variance sequence output by the feature extraction and demodulation module, and performs threshold comparison in conjunction with the backscattering coefficient to generate a weighted mask sequence for characterizing the abnormal state of the spatial node.

[0026] The noise reduction module simultaneously receives the scattering intensity variance sequence output by the feature extraction and demodulation module and the original spatial phase signal. The noise reduction module runs a Kalman filter algorithm, maps the scattering intensity variance sequence as an input variable to the state space model of the Kalman filter, updates the measurement noise covariance matrix of the algorithm node by node in the spatial dimension, and performs a smoothing operation on the original spatial phase signal based on the updated matrix parameters, and outputs a filtered phase signal.

[0027] The compensation and reconstruction module receives the attenuation coefficient and weighted mask sequence output from the optical domain calibration module, as well as the filtered phase signal output from the noise reduction module. It calculates the spatial baseline offset using the attenuation coefficient and generates a weighted theoretical baseline drift term by combining it with the weighted mask sequence. The module then subtracts the weighted theoretical baseline drift term from the filtered phase signal to obtain the baseline compensation sequence. Finally, it performs continuous unwrapping on the baseline compensation sequence and reconstructs the final acoustic vibration signal based on a mapping model.

[0028] In this low-temperature crosstalk distributed acoustic sensing system, the internal light guiding medium of the anti-resonant hollow-core sensing fiber is air. The total optical phase change caused by external stress acting on the fiber consists of two parts: the first part is the change of the fiber core equivalent refractive index with pressure; the second part is the change of the fiber axial geometric length with pressure.

[0029] Because the refractive index of the air medium inside the sealed anti-resonant hollow-core sensing fiber is constant, its equivalent refractive index response to external pressure approaches zero. Therefore, the phase response induced by external strain modulation consists only of the axial geometric deformation term of the fiber. This eliminates the influence of the elasto-optic effect of the medium material itself and achieves parametric decoupling between the strain phase and temperature phase at the signal excitation level, providing the subsequent processing module with the original parameters least affected by temperature crosstalk.

[0030] Reference Figure 1 The optical signal module of the present invention is configured to generate a detection signal and complete the layer-level sensing interaction and heterodyne coherent detection, including a transmitting excitation unit, a sensing medium unit and a receiving mixing unit.

[0031] The excitation unit comprises a narrow-linewidth laser, an optical splitter, a pulse modulation module, and a power amplifier. The narrow-linewidth laser generates a continuous optical signal with a fixed center wavelength, which is transmitted to the optical splitter via a single-mode fiber. The optical splitter divides the continuous optical signal into a local oscillator reference beam and a primary probe beam according to a set splitting ratio. The primary probe beam is input to the pulse modulation module, which typically employs an acousto-optic modulator. This modulator modulates the amplitude of the primary probe beam to generate a probe beam pulse with a specific pulse width, and simultaneously applies a fixed acousto-optic frequency shift. The modulated probe beam pulse is then input to the power amplifier for optical power amplification to compensate for transmission losses in subsequent links.

[0032] The sensing medium unit comprises a circulator and an anti-resonant hollow-core sensing fiber sealed at both ends. A probe light pulse, amplified by a power amplifier, enters the first port of the circulator and exits from the second port, then is injected into the anti-resonant hollow-core sensing fiber and propagates axially. When the anti-resonant hollow-core sensing fiber is installed in the environment under test, external strain pressure applied to it causes a change in the optical phase.

[0033] In terms of sensing mechanism, external strain pressure Acting on spatial position When the optical fiber is subjected to external pressure, the relationship between the resulting optical phase change and the external pressure is determined by the fiber strain sensitivity coefficient. The definition, its mathematical expression is: ; In the formula, The fiber strain sensitivity coefficient; For light phase; The equivalent refractive index of the fundamental mode of the anti-resonant hollow-core sensing fiber; The length of the optical fiber in the strained region; This represents the strain pressure exerted by external forces. Because the sensing medium employs a sealed structure at both ends, the equivalent refractive index of the air inside the anti-resonant hollow-core sensing fiber remains constant with external pressure, leading to a change in the derivative term of the equivalent refractive index. The value approaches zero. Therefore, the phase modulation in the sensing medium unit consists only of the axial geometric deformation term. constitute.

[0034] Simultaneously, the fiber temperature sensitivity coefficient of the anti-resonant hollow-core sensing fiber... Controlled by the thermal expansion characteristics of its multilayer material structure, its mathematical expression is: ; In the formula, The temperature sensitivity coefficient of the optical fiber; For temperature parameters; The length of the optical fiber in the strained region; Indicates the fiber optic cable number 1 The coefficient of thermal expansion of the layer; Indicates the fiber optic cable number 1 Young's modulus of the layer; Indicates the fiber optic cable number 1 The cross-sectional area of ​​the layer. Combining strain and temperature response parameters, the optical signal module, during the material selection phase, limits the structural parameters of the multilayer material based on the acoustic-temperature quality factor. The theoretical evaluation criteria are established, and their relationship is as follows: ; In the formula, For sound temperature quality factor; The fiber strain sensitivity coefficient; This is the temperature sensitivity coefficient of the optical fiber.

[0035] The receiving mixer unit includes a preamplifier, an optical bandpass filter, and a coherent detection module. Backscattered Rayleigh light carrying phase information, generated within the anti-resonant hollow-core sensing fiber, returns along its original path, enters the second port of the circulator, and is output from the third port. The backscattered Rayleigh light is input to the preamplifier for power amplification at the receiving end. The amplified optical signal then passes through the optical bandpass filter, which filters out out-of-band spontaneous emission amplification noise.

[0036] The filtered backscattered Rayleigh light and the local oscillator reference light reserved in the emission excitation unit are simultaneously input into the coherent detection module. The coherent detection module uses a photoelectric balanced detector to perform optical interference mixing between the local oscillator reference light with the backscattered Rayleigh light, which has a frequency difference, and converts the mixed optical signal into an analog electrical signal. The analog electrical signal undergoes analog-to-digital conversion and is finally output as an interference electrical signal to the feature extraction and demodulation module.

[0037] Reference Figure 1 The present invention provides a feature extraction and demodulation module configured to receive the interference electrical signal output by the optical signal module and perform parametric feature separation and spatial phase operation, including an envelope extraction unit, a variance calculation unit and a spatial phase demodulation unit.

[0038] The envelope extraction unit is connected to the receiving mixer unit to receive the mixed interference signal. Since the interference signal carries an intermediate frequency (IF) carrier component generated by coherent detection, the envelope extraction unit performs orthogonal down-conversion and low-pass filtering on the interference signal to remove high-frequency and IF carrier components, obtaining mutually orthogonal in-phase and quadrature components. The envelope extraction unit calculates the sum of squares of the in-phase and quadrature components and performs a square root operation on the sum of squares to extract the backscattered light intensity envelope data sequence corresponding to each discrete spatially distributed node of the anti-resonant hollow-core sensing fiber.

[0039] The variance calculation unit is connected to the envelope extraction unit and receives the backscattered light intensity envelope data sequence. Due to the weak backscattering characteristics of the air medium inside the anti-resonant hollow-core sensing fiber, deformation of the local spatial structure will cause random fluctuations in light intensity. The variance calculation unit first sets a fixed preset time window on the discrete time axis, and calculates the time mean of the backscattered light intensity envelope by summing all the backscattered light intensity envelope sample values ​​contained within the preset time window and dividing by the total number of sampling points.

[0040] Subsequently, the variance calculation unit calculates the dispersion of signal fluctuations based on the time mean, generating a scattering intensity variance sequence. Specific characteristic statistical parameters are calculated based on the following mathematical equations: ; In the formula, Representing spatial nodes The variance of scattering intensity at a given location is formed by splicing together the calculated values ​​of each spatial node to create the overall scattering intensity variance sequence. This indicates the total number of discrete sampling points included within the preset time window; Representing spatial nodes In the The backscattered light intensity envelope value corresponding to each sampling time; This represents the time mean of the backscattered light intensity envelope corresponding to the spatial node within a preset time window; This parameter represents the index of the sampling points in a discrete-time series.

[0041] The spatial phase demodulation unit is connected to the optical signal module and configured to process the interferometric electrical signal and extract the phase parameter applied by external strain. The spatial phase demodulation unit receives the interferometric electrical signal and uses a hardware 3x3 symmetric coupler or a digitally equivalent phase-splitting algorithm to separate the interferometric electrical signal into three voltage signals with a nominal phase difference of 120 degrees between them.

[0042] Because the three voltage signals carry DC bias parameters and amplitude asymmetry caused by gain differences during device conversion, the spatial phase demodulation unit performs averaging and subtraction operations on the three voltage signals to remove the DC bias and performs amplitude normalization. The spatial phase demodulation unit performs nonlinear combination operations on the normalized three voltage signals according to the arctangent demodulation equation, the mathematical expression of which is as follows: ; In the formula, Representing spatial nodes exist The original spatial phase signal demodulated at each moment; This represents the first voltage signal after normalization. This represents the arctangent mathematical operation function; Represents the constant term, i.e. The square root of; This represents the second voltage signal after normalization. This represents the third voltage signal after normalization. After performing the above calculations, the spatial phase demodulation unit unwraps the carrier phase and outputs the original spatial phase signal, which has a two-dimensional distribution in both time and space.

[0043] Reference Figure 1 The optical domain calibration module of the present invention is configured to independently measure the optical parameters along the anti-resonant hollow core sensing fiber, and generate spatial shielding parameters by combining the dynamic statistical features output by the feature extraction and demodulation module, including a parameter measurement separation unit, an abnormal node identification unit and a mask generation unit.

[0044] The parameter measurement separation unit is connected to both ends of the anti-resonant hollow-core sensing fiber. The backscattered signal obtained by conventional unidirectional optical time-domain reflectometry (OTDR) is the coupling result of the local backscattering coefficient and the bidirectional attenuation along the transmission path. The parameter measurement separation unit executes a bidirectional ODR calibration procedure. First, a forward test pulse is emitted from the first end of the anti-resonant hollow-core sensing fiber, and the forward optical power distribution curve returned along the path is recorded. Then, the hardware optical path is switched, and a reverse test pulse is emitted from the second end of the anti-resonant hollow-core sensing fiber, recording the reverse optical power distribution curve returned along the path.

[0045] The parameter measurement and separation unit separates the attenuation coefficient at the same spatial node by performing logarithmic subtraction and differentiation on the forward and reverse optical power distribution curves. Simultaneously, the unit performs a geometrical averaging of the values ​​from the forward and reverse optical power distribution curves to eliminate the coupling effect of the attenuation term along the path, thus separating the backscattering coefficient directly related to the local medium structure. The mathematical equation for the backscattering coefficient is as follows: ; In the formula, Indicates the spatial nodes obtained by separation The backscattering coefficient at that location; This represents the system constant, which is determined by the initial transmit power of the bidirectional test pulse and the fixed insertion loss at both ends of the optical fiber. Indicates at spatial nodes The numerical values ​​of the forward optical power distribution recorded at that location; Indicates the same spatial node The recorded values ​​of the reverse light power distribution.

[0046] The anomaly node identification unit is connected to both the parameter measurement and separation unit and the front-end feature extraction and demodulation module, receiving the backscattering coefficient and scattering intensity variance sequences distributed in the spatial domain. When the anti-resonant hollow-core sensing fiber is subjected to non-axial anomalous compression, its internal microstructure tubes undergo lateral deformation, leading to a change in the intrinsic scattering cross-section at that location and generating excessive scattering intensity fluctuations. The anomaly node identification unit aligns the backscattering coefficient and scattering intensity variance sequences according to the spatial distance coordinates of the anti-resonant hollow-core sensing fiber, obtaining the numerical state of the backscattering coefficient and scattering intensity variance sequences corresponding to each spatial node.

[0047] The abnormal node identification unit internally stores pre-set backscattering anomaly thresholds and intensity variance thresholds. The unit iterates through all spatial nodes, comparing the backscattering coefficient of each node with the backscattering anomaly threshold, and comparing the scattering intensity variance sequence value of each node with the intensity variance threshold. When the backscattering coefficient of a spatial node is determined to be greater than the backscattering anomaly threshold, or the scattering intensity variance sequence value of that spatial node is determined to be greater than the intensity variance threshold, the abnormal node identification unit marks that spatial node as a nonlinear interference node.

[0048] The mask generation unit receives the node label status output by the abnormal node identification unit and assigns a specific numerical weight factor to each spatial node based on the numerical status. The mask generation unit executes the allocation logic based on a piecewise function, the mathematical expression of which is as follows: ; In the formula, Representing spatial nodes The weight factor values ​​generated at that location; This represents the set of spatial coordinates of nodes marked as nonlinear interference nodes. Belongs to the symbol category, representing spatial nodes. It belongs to the set of nonlinear interference nodes; It is not a symbol, but represents a spatial node. It does not belong to the set of nonlinear interference nodes, meaning that the node is in a normal state; Represents spatial nodes; This represents the preset decay penalty factor. The value range of is strictly limited to the interval between greater than zero and less than one. The mask generation unit concatenates the weight factor values ​​of all spatial nodes in spatial coordinate order to form a one-dimensional weight mask sequence, and outputs it to the subsequent module to participate in the numerical constraint of the baseline term.

[0049] Reference Figure 1 The noise reduction module of the present invention is configured to receive the original spatial phase signal and the scattering intensity variance sequence, and perform a smoothing operation to suppress local random abrupt noise in space by driving the upper-level digital algorithm through the underlying parameters. It includes a state parameter mapping unit, a covariance update unit and a filtering execution unit.

[0050] To address the backscattering signal fading phenomenon in anti-resonant hollow-core sensing fibers caused by local structural perturbations, conventional Kalman filtering algorithms, employing a fixed measurement noise covariance matrix, cannot adapt to non-uniformly distributed random noise in the spatial domain. A state parameter mapping unit establishes a numerical relationship between physical characteristics and algorithm parameters. It receives the scattering intensity variance sequence from the previous stage and extracts the spatial node values, which exhibit a discrete spatial distribution. The state parameter mapping unit multiplies the extracted spatial node values ​​with a preset measurement noise reference parameter to obtain the numerical amplification factor for each spatial node. This numerical amplification factor is then directly output as a covariance adjustment parameter. The mathematical expression for this mapping process is as follows: ; In the formula, Representing spatial nodes The covariance adjustment parameter generated at the location; This represents the variance sequence of scattering intensity at spatial nodes. The input node value at the location; This represents the initial measurement noise reference parameter set by the system.

[0051] The covariance update unit is connected to the state parameter mapping unit and receives the covariance adjustment parameter. The state-space model of the Kalman filter includes the process noise covariance matrix and the measurement noise covariance matrix. The covariance update unit substitutes the covariance adjustment parameter into the Kalman filter model and, according to the node coordinates in the spatial dimension, replaces the row and column elements of the measurement noise covariance matrix node by node. When the scattering intensity variance of a spatial node increases, the replaced measurement noise covariance matrix value increases proportionally, completing the adaptive spatial value replacement.

[0052] The filtering execution unit is connected to the covariance update unit and the feature extraction and demodulation module, and reads the updated measurement noise covariance matrix and the input original spatial phase signal. The filtering execution unit first calculates the Kalman gain based on the prior estimation error covariance matrix and the updated measurement noise covariance matrix. The mathematical expression of its Kalman gain equation is: ; In the formula, Representing spatial nodes exist Kalman gain matrix at time step; The prior estimate error covariance matrix represents the state variable; Represents the observation matrix; This represents the transpose of the observation matrix; This represents the updated measurement noise covariance matrix, whose value is equal to the covariance adjustment parameter. .

[0053] According to the matrix inversion logic of the Kalman gain equation, when the updated measurement noise covariance matrix... As the value of increases, the total number of inverse terms increases, directly affecting the Kalman gain of the calculated output. The value decreased.

[0054] Subsequently, the filtering execution unit substitutes the calculated Kalman gain into the posterior state update equation to perform smoothing processing on the original spatial phase signal. The mathematical expression of its state update equation is as follows: ; In the formula, This represents the posterior state estimate after smoothing, i.e., the filtered phase signal output by the filtering execution unit; This represents the prior state prediction value derived by the system based on the state at the previous time step; This represents the current observation value, i.e., the original spatial phase signal input from the previous stage; Representing spatial nodes exist Kalman gain matrix at time step; This represents the observation matrix.

[0055] When distortion occurs at spatial nodes due to external disturbances, resulting in a large intensity variance, the Kalman gain decreases synchronously. It will reduce the observation value containing mutation noise at the current time. The numerical weighting in the state update equation forces the filter output to be more inclined towards the prior state prediction. Through closed-loop linkage calculation of this parameter and the digital filtering matrix variables, the filtering execution unit completes spatial point-by-point adaptive noise reduction of the original spatial phase signal and finally outputs the filtered phase signal.

[0056] Reference Figure 1 The compensation and reconstruction module of the present invention is configured to eliminate the system baseline offset caused by long-distance transmission and convert the clean phase data into an external parameter matrix, including a drift modeling unit, an interpolation compensation unit and an unwrapping reconstruction unit.

[0057] The drift modeling unit is connected to both the noise reduction module and the optical domain calibration module, receiving the attenuation coefficient and weight mask sequence. In distributed long-distance sensing links, the intrinsic attenuation of the optical fiber and local physical deformation cause a cumulative DC baseline shift in the demodulated phase along the spatial distance. The drift modeling unit first performs spatial domain integration calculations on the discrete spatially distributed attenuation coefficients combined with photoelectric conversion parameters to generate the initial spatial phase shift.

[0058] Subsequently, the drift modeling unit multiplies the weighted mask sequence with the initial spatial phase offset according to the coordinate positions corresponding to the spatial nodes. This multiplication operation uses the weighting factor output by the mask generation unit to proportionally reduce the spatial phase offset value of the regions marked as nonlinear interference nodes, ultimately outputting a weighted theoretical baseline drift term. The mathematical expression for this process is as follows: ; In the formula, This represents the spatial node. The corresponding weighted theoretical baseline drift term; Represents the corresponding spatial node The weight mask value at the location; This represents the conversion constant from a fixed decay to a phase. Indicates the first The attenuation coefficients measured separately at each spatial node; Indicates the spatial sampling interval; This represents the total number of indices of spatial nodes, corresponding to the current spatial position as the calculation ends. This represents the discrete node index variable in the spatial domain cumulative calculation.

[0059] The differential compensation unit is connected to the drift modeling unit, receiving the filtered phase signal and the weighted theoretical baseline drift term. On the two-dimensional matrix plane of time and space, the differential compensation unit reads the value of the filtered phase signal node by node and subtracts the weighted theoretical baseline drift term for the corresponding spatial location from this value. This subtraction operation eliminates the static phase shift caused by optical path loss and non-axial compression, ensuring that the phase reference of all spatial nodes is zeroed, and outputting the baseline-flattened compensation sequence. Its mathematical expression is: ; In the formula, Representing spatial nodes exist The baseline compensation sequence generated at each time step; This represents the posterior state estimate after smoothing, i.e., the filtered phase signal output by the filtering execution unit; This represents the spatial node. The corresponding weighted theoretical baseline drift term.

[0060] The unwrapping and reconstruction unit is connected to the difference compensation unit and receives the baseline compensation sequence. Due to the computational limitations of the preceding arctangent mathematical algorithm, the input baseline compensation sequence values ​​are confined within the range of negative to positive pi, exhibiting jagged, discontinuous jumps. The unwrapping and reconstruction unit performs a branch-cutting algorithm on the two-dimensional distributed baseline compensation sequence. By identifying residual charge points in the phase matrix, a cutting line is established between the positive and negative residual poles. Subsequently, phase difference accumulation operations are performed on adjacent pixels along the path bypassing the cutting line. After processing by this unwrapping algorithm, the portion of the baseline compensation sequence with jump values ​​equal to or close to twice pi is restored, outputting continuous phase data that is continuously distributed on the time axis.

[0061] Finally, the unwrapping and reconstruction unit reads the strain sensitivity coefficient of the aforementioned defined anti-resonant hollow-core sensing fiber and constructs a pure geometric phase stress inverse mapping model. The unwrapping and reconstruction unit substitutes the continuous phase data obtained from the unwrapping into this model to calculate the amplitude parameter of the acoustic vibration signal in reverse. Its mathematical mapping equation is as follows: ; In the formula, This represents the amplitude of the acoustic vibration signal at the final reconstructed output; This represents the continuous phase data obtained from unwrapping; The strain sensitivity coefficient of the anti-resonant hollow-core sensing fiber is represented by this value. This represents the optical reference phase of the initially injected probe pulse; The length of the optical fiber in the strained region is denoted as . Through the above calculations, the unwrapping and reconstruction unit completes the parametric conversion and reconstruction output of the optical signal into an acoustic vibration signal.

[0062] Specific application examples: In an experimental environment simulating long-distance oil and gas pipeline monitoring, the anti-resonant hollow-core sensing fiber of this invention and a traditional commercial single-mode solid-core fiber were laid parallel to each other in the same trench on the outer wall of the test pipeline. The total length of the sensing fiber was set to 10 kilometers. Two independent variables were introduced in the experiment: Low-frequency strain excitation: A continuous sinusoidal low-frequency strain signal with a frequency of 1 Hz and an equivalent micro-strain of 50 με50 με is applied at a distance of 5 km from the beginning of the optical fiber using a piezoelectric ceramic resonator ring to simulate the low-frequency vibration of the pipe wall caused by a small leak in the pipeline.

[0063] Temperature interference application: In a 200-meter fiber optic segment from 4.9 km to 5.1 km, a programmable temperature control box is used to apply slow temperature fluctuations with a period of 50 seconds, linearly increasing from 20°C to 60°C and then decreasing, to simulate low-frequency temperature interference from day-night temperature differences or external environment.

[0064] The optical signal module injects probe light pulses with identical parameters into the two links respectively; the feature extraction and demodulation module processes the returned interference electrical signals.

[0065] The first comparison focuses on the effect of suppressing temperature crosstalk. Traditional solid-core optical fibers are affected by the thermo-elastic-optical effect of silica, making their refractive index highly sensitive to temperature changes. In the phase signal demodulated from traditional solid-core optical fibers, a temperature drift of 0.02 Hz (corresponding to a 50-second period) causes a baseline shift of over 80 rad, completely obscuring the 1 Hz true strain signal and easily leading to cycle skipping errors in subsequent unwrapping algorithms. This invention uses an anti-resonant hollow-core sensing fiber sealed at both ends. Because its internal light-guiding medium is air, its equivalent refractive index has low temperature sensitivity. Under the same temperature change conditions, the baseline drift of the original spatial phase signal demodulated at the corresponding spatial node in this invention's system is controlled within 2 rad, and the 1 Hz target strain characteristics remain clearly visible. This experimental data verifies the physical decoupling capability of the sensing medium of this invention against low-frequency temperature crosstalk.

[0066] The second comparison focuses on the suppression effect of local random abrupt noise. In hollow-core fiber applications, random compression of local capillaries causes severe fading of the Rayleigh scattering signal at spatial nodes, manifesting as sudden spike noise in the demodulated phase. When the noise reduction module is off, random spikes with amplitudes as high as 15 rad exist in the original spatial phase signal, severely disrupting signal continuity. After the noise reduction module is on, the state parameter mapping unit detects the abnormal increase in the variance sequence of the scattering intensity of the spatial node, instantly generating an amplified covariance adjustment parameter to update the measurement noise covariance matrix. The Kalman gain is adaptively reduced, and the system no longer trusts the observed value at that moment but tends to rely on the prior prediction value. The final acoustic vibration signal output by the compensation and reconstruction module is smooth and continuous, with the amplitude of sudden spike noise reduced by more than 90%, while the 1Hz target signal waveform remains undistorted and unattenuated.

[0067] To visually demonstrate the verification process described above, the following two test data waveforms are provided for comparison: Reference Figure 2A comparison of the phase responses of different sensing media under low-frequency temperature interference. Figure 2 The baseline drift states of the demodulated phase of conventional solid fiber and the demodulated phase of the anti-resonant hollow fiber of this invention are shown under the same 1Hz strain and severe temperature fluctuation.

[0068] Reference Figure 3 The image shows a comparison of the phase waveforms before and after processing by the noise reduction module of this invention. Figure 3 The waveforms of the original spatial phase signal with scattering fading abrupt noise and the filtered phase signal after being processed by the variance-adaptive Kalman filter of this invention are shown at the same node.

Claims

1. A low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber, characterized in that, include: The optical signal module is used to inject the probe light pulse into the anti-resonant hollow core sensing fiber, receive the returned back Rayleigh scattered light and heterodyne mix it with the local oscillator reference light to output an interference electrical signal. The feature extraction and demodulation module is used to extract the backscattered light intensity envelope from the interference electrical signal, calculate the time variance to generate a scattering intensity variance sequence, and demodulate the original spatial phase signal from the interference electrical signal. The optical domain calibration module is used to separate the attenuation coefficient and backscattering coefficient of the anti-resonant hollow-core sensing fiber, and to generate a weighted mask sequence by combining the backscattering coefficient and the scattering intensity variance sequence. The noise reduction module is used to reduce the noise of the original spatial phase signal using Kalman filtering, map and update the measurement noise covariance matrix using the scattering intensity variance sequence, and output the filtered phase signal. The compensation and reconstruction module is used to establish a weighted theoretical baseline drift term using the attenuation coefficient and the weighted mask sequence, subtract the weighted theoretical baseline drift term from the filtered phase signal to obtain a baseline compensation sequence, and unwrap the baseline compensation sequence to reconstruct the acoustic vibration signal.

2. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 1, characterized in that, The optical signal module includes: The emission excitation unit is used to generate continuous light and modulate the continuous light into the probe light pulse, amplify the probe light pulse, and inject it into the anti-resonant hollow core sensing fiber. A sensing medium unit is used to conduct the probe light pulse, which undergoes axial geometric deformation when modulated by external strain and returns the back Rayleigh scattered light carrying axial geometric deformation information. A receiving mixer unit is used to receive the backscattered Rayleigh light and amplify the signal, perform heterodyne mixing operation between the amplified backscattered Rayleigh light and the local oscillator reference light, and output the interference electrical signal.

3. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 1, characterized in that, The feature extraction and demodulation module includes: An envelope extraction unit is used to receive the interference electrical signal and separate the interference electrical signal to obtain the backscattered light intensity envelope corresponding to the spatial distribution node of the anti-resonant hollow sensing fiber. The variance calculation unit is used to obtain the time mean of the backscattered light intensity envelope within a time window, calculate the squared difference between the backscattered light intensity envelope and the time mean, and combine the values ​​to generate the scattering intensity variance sequence. The spatial phase demodulation unit is used to perform a phase demodulation mathematical algorithm on the interference electrical signal, extract and output the original spatial phase signal.

4. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 3, characterized in that, The spatial phase demodulation unit is used to receive the interference electrical signal, separate the interference electrical signal into three voltage signals with a phase difference of 120° between them, perform mathematical calculations on the three voltage signals with a phase difference of 120° between them according to the arctangent demodulation equation, and output the original spatial phase signal.

5. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 1, characterized in that, The optical domain calibration module includes: The parameter measurement and separation unit is used to transmit test pulses to the anti-resonant hollow-core sensing fiber, record the optical power distribution curve along the return path, and obtain the attenuation coefficient and the backscattering coefficient by calculating the optical power distribution curve. An abnormal node identification unit is used to perform data mapping and alignment between the backscattering coefficient and the scattering intensity variance sequence to obtain the numerical status of the backscattering coefficient and scattering intensity variance sequence of the spatial node. The mask generation unit is used to assign weight factors to the spatial nodes according to the numerical state, and to concatenate the weight factors of all the spatial nodes to form the weight mask sequence.

6. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 5, characterized in that, The abnormal node identification unit is used to receive a backscattering anomaly threshold and an intensity variance threshold, and to perform a comparison operation between the backscattering coefficient and the backscattering anomaly threshold, and to perform a comparison operation between the value of the corresponding spatial node in the scattering intensity variance sequence and the intensity variance threshold. When the backscattering coefficient is greater than the backscattering anomaly threshold or the value of the corresponding spatial node is greater than the intensity variance threshold, the corresponding spatial node is marked as a nonlinear interference node.

7. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 1, characterized in that, The noise reduction module includes: The state parameter mapping unit is used to receive the scattering intensity variance sequence and input the scattering intensity variance sequence into the state space model mapping function to generate covariance adjustment parameters. The covariance update unit is used to substitute the covariance adjustment parameter into the Kalman filter and replace the row and column values ​​of the measurement noise covariance matrix node by node in the spatial dimension. The filtering execution unit is used to read the updated measurement noise covariance matrix, calculate the Kalman gain, substitute the calculated Kalman gain into the posterior state update equation, and perform smoothing processing on the original spatial phase signal.

8. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 7, characterized in that, The state parameter mapping unit is used to extract the spatial node values ​​in the scattering intensity variance sequence, multiply the spatial node values ​​with the measurement noise reference parameter to obtain the numerical amplification factor, and output the numerical amplification factor as the covariance adjustment parameter.

9. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 1, characterized in that, The compensation and reconstruction module includes: A drift modeling unit is used to calculate the spatial phase offset using the attenuation coefficient, multiply the weighted mask sequence with the spatial phase offset, and output the weighted theoretical baseline drift term. The difference compensation unit is used to receive the filtered phase signal and the weighted theoretical baseline drift term, perform a subtraction operation on the filtered phase signal and the weighted theoretical baseline drift term, and output the baseline compensation sequence. The untangling and reconstruction unit is used to perform a phase unwinding algorithm on the baseline compensation sequence to remove the tangled phase, and then combine the pure geometric phase stress inverse mapping model to convert the baseline compensation sequence with the tangled phase removed into the acoustic vibration signal.

10. The low-temperature crosstalk distributed acoustic sensing system based on anti-resonant hollow optical fiber according to claim 9, characterized in that, The unwrapping and reconstruction unit is used to perform a branch cutting algorithm on the baseline compensation sequence, expand the baseline compensation sequence wrapped in the range of negative to positive pi values ​​to obtain continuous phase data, substitute the continuous phase data into the pure geometric phase stress inverse mapping model, reverse solve and output the amplitude and frequency parameters of the acoustic vibration signal.

Images