WO2020181920A1

WO2020181920A1 - Signal processing method based on distributed fiber-optic acoustic sensor

Info

Publication number: WO2020181920A1
Application number: PCT/CN2020/072557
Authority: WO
Inventors: 蔡海文; 梁嘉靖; 郑汉荣; 王照勇; 卢斌; 叶青
Original assignee: 中国科学院上海光学精密机械研究所
Priority date: 2019-03-14
Filing date: 2020-01-16
Publication date: 2020-09-17
Also published as: CN111693131A

Abstract

Provided is a signal processing method based on a distributed fiber-optic acoustic sensor; in said method, spatial position information of a signal source is obtained by means of pre-processing an acoustic field signal obtained by a sensing unit; then signal processing is performed on the described information, effectively suppressing system random noise introduced by the distributed fiber-optic acoustic sensing system itself, increasing system sensitivity, while also having the ability to enhance or suppress specific incoming and specific spatial position signals, enabling directional and fixed-point detection of perturbation signals. The present invention has advantages such as simple implementation, fast processing speed, strong interference-resistance capability, and obvious improvement of signal-to-noise ratio; further improves the ability of existing sensing systems to monitor target interference signals in a complex operating environment; is suitable for use in railway safety, oil and gas pipeline monitoring, perimeter security, and the like; and has great significance.

Description

A signal processing method based on distributed optical fiber acoustic sensor

Technical field

The invention relates to the field of signal source monitoring, in particular to a signal processing method based on a distributed optical fiber acoustic sensor.

Background technique

Distributed optical fiber acoustic sensors are widely used in railway safety, oil and gas pipeline monitoring, perimeter security and other fields. Most of the existing distributed optical fiber acoustic sensors use ordinary communication optical fibers as the sensing fibers. On the one hand, they have low transduction coefficients for disturbance signals in the application scenarios of distributed optical fiber acoustic sensors, and their ability to detect weak signals is limited by the sensor. The base noise of the system; on the other hand, the application scenarios of distributed optical fiber acoustic sensors often have environmental noise and interference signals other than the target signal, making it difficult for the sensing system to effectively identify the target signal in a complex working environment.

Prior art one [Pan Z, Cai H, Qu R, et al. Phase-sensitive OTDR system based on digital coherent detection. Asia Communications & Photonics Conference & Exhibition. IEEE, 2012.] Phase-sensitive demodulation based on digital coherent demodulation is proposed. The sensitive optical time domain reflectometer quantifies the measurement system and the demodulation formula of amplitude and phase information, but it does not further use the spatial correlation between the quantified sound field signals to enhance the sound field signals.

Existing technology two [Yang G, Fan X, Wang S, et al. Long-Range Distributed Vibration Sensing Based on Phase Extraction From Phase-Sensitive OTDR. IEEE Photonics Journal, 2016.] proposed a compensation for laser phase noise Method to improve the signal-to-noise ratio of the phase-sensitive optical time domain reflectometer under long-distance monitoring, but this method needs to set weak reflection points along the sensing fiber to extract the laser phase noise.

Existing technology three [Martins H, Shi K, Thomsen B, et al. Real time dynamic strain monitoring of optical links using the back reflection of live PSK data[J]. Optics Express, 2016.] Utilizes pulse coding technology to be sensitive to phase The sensing signal of the optical time domain reflectometer is enhanced, which effectively improves the signal-to-noise ratio, but this method does not have the directional detection capability, and it is difficult to suppress the interference signal.

Existing technology four [Liu T, Wang F, Zhang X, et al. Phase sensitive distributed vibration sensing based on ultrarawak fiber Bragg grating array using double-pulse [J]. Optical Engineering, 2017.] The enhancement of the backscattered light of the optical fiber can effectively suppress the Rayleigh coherent noise in the phase-sensitive optical time domain reflectometer and improve the signal-to-noise ratio of the disturbance signal detection. However, this method does not have the directional detection capability and is difficult to perform for the disturbance signal. inhibition.

Summary of the invention

In order to overcome the above-mentioned shortcomings of the prior art, the present invention proposes a signal processing method based on distributed optical fiber acoustic sensors, which enables the distributed sensing system to detect disturbance signals in a directional and fixed-point manner, and at the same time has the ability to enhance and suppress specific directions or specific spaces. The ability of location signals. This method is suitable for existing sensing systems and has the advantages of simple implementation, fast processing speed, strong anti-interference ability, and obvious improvement in signal-to-noise ratio. It can greatly improve the use of distributed optical fiber acoustic sensors in railway safety, oil and gas pipeline monitoring, and perimeter The ability to monitor target signals in complex working environments such as security.

The present invention proposes a signal processing method based on a distributed optical fiber acoustic sensor, which includes a signal source, a distributed optical fiber acoustic sensor, and a sensing optical fiber. The method is characterized in that the method is as shown in Figure 1 and includes the following steps:

1) The distributed optical fiber acoustic sensor transmits detection light pulses to the sensing fiber, quantitatively monitors the sound field sensed along the sensing fiber, and obtains the sound field distribution signal S(l, t), where l Represents the one-dimensional axial space of the sensing fiber, and t is time;

2) The distributed optical fiber acoustic sensor preprocesses the obtained sound field signal:

First, select a section of length L in the sensing fiber as the enhanced aperture, and calculate the spatial position information of the signal received in the enhanced aperture;

Secondly, the number of sensing units included in the enhanced aperture is N, which is determined by the spatial sampling frequency f _{s of the} distributed optical fiber acoustic sensor, which is taken at intervals of Δn,

The sound field signals collected by two sensing units are used as the sampling signal group,

Among them, l ₁ represents the starting position of the enhanced aperture on the fiber link;

Finally, according to the spatial position information, calculate the time delay vector of the signal source with respect to each sensing unit in the enhanced aperture,

Or phase delay vector

Where m=1, 2,...,M, M represents the number of signal sources;

3) Use the array signal processing method to perform signal processing on the sampled signal group and delay vector obtained in step 2).

In the step 2), the spatial position information is the azimuth vector of the signal source relative to the enhanced aperture or the three-dimensional position coordinates of the signal source relative to the enhanced aperture.

In the step 2), the spatial position information is obtained by using the signal source spatial positioning method or artificially designated according to prior knowledge.

The interval of the enhanced aperture is determined according to the requirements of specific application scenarios, and the calculation time and the enhancement effect need to be weighed. The sampling signal group includes the number of sensing units according to the requirements of specific application scenarios, and the calculation time and the enhancement effect need to be weighed. The signal source spatial positioning method based on distributed optical fiber sensing does not belong to the scope of this patent. The spatial information may be only the azimuth vector of the signal source relative to the enhanced aperture, or only the three-dimensional position coordinates of the signal source relative to the enhanced aperture, or both may be included, depending on actual application scenarios and application requirements.

The array signal processing method in the step 3) is an adaptive spatial filtering method or a delay sum method.

The adaptive spatial filtering method is one of a minimum variance distortionless response beamformer (MVDR), a linearly constrained minimum variance beamformer (LCMV), and a generalized sidelobe canceling beamformer (GSC).

The delay summation method is as follows:

Calculate the delay compensation weight,

For the m-th signal source, its enhanced signal is calculated in the following way,

For other M-1 signal sources, the signal suppression effect is limited by the size of the enhanced aperture. If the m'th signal source needs to be enhanced, recalculate its delay compensation weight w _m' (Δτ), and calculate the enhanced signal according to the above equation Just output Y _m' (t).

The linear constrained minimum variance beamformer (LCMV) method is as follows:

Calculate the covariance matrix of the sampled signal group,

Wherein K represents the number of repetitions of the distributed optical fiber acoustic sensor emitting detection light pulses to the sensing optical fiber;

For the m-th signal source, determine the restriction vector F _1×M = [0 0 ... 1 ... 0], where the m-th element of the restriction vector is 1, and the rest are 0. The restriction vector determines that only The m-th signal source performs directional enhancement, while the other signal sources perform directional suppression. The phase compensation weight is calculated according to the following equation,

Among them, H represents conjugate transpose,

For the m-th signal source, the enhanced signal is calculated according to the following equations,

To enhance the m'th signal source, modify the restriction vector F _1×M and recalculate the phase compensation weight

And calculate the output Y _m' (t) of the enhanced signal according to the above equation.

When the signal source is a broadband signal, the step 2) further includes:

Transform the sampling signal group to the frequency domain sampling signal group X(f), and split it into P subband signals X(f _p ),

According to the spatial position information of the signal source relative to the enhanced aperture, calculate the phase delay vector of each subband of the signal source with respect to each sensing unit

Or time delay vector

The step 3) is to use a linearly constrained minimum variance beamformer to process the frequency domain sampled signal group, specifically:

First calculate the covariance matrix corresponding to each sub-band,

Where J represents the frequency accuracy of the fast Fourier transform of the distributed optical fiber acoustic sensor on the time-domain sampling signal group;

Secondly, for the m-th signal source, determine the restriction vector F _1×M = [0 0 ... 1 ... 0], where the m-th element of the restriction vector is 1, and the rest are 0. The restriction vector is determined to be only The m-th signal source is directional enhancement, and the other signal sources are directional suppression. The phase compensation weight corresponding to each sub-band is calculated according to the following equation,

Among them, H represents conjugate transpose,

Finally, for the p-th subband of the m-th signal source, the frequency-domain enhanced signal is calculated according to the following equation,

All the subbands are reconstructed into a complete frequency domain signal, and then through the inverse fast Fourier transform, the time domain signal of the m-th signal source after enhancement can be obtained. If it is necessary to enhance the m'th signal source, modify the restriction vector F _1×M and recalculate the phase compensation weight of each subband

And calculate the frequency domain enhanced signal of each subband according to the above equation

OK.

The features and advantages of the present invention are:

1. The present invention is suitable for the existing distributed optical fiber acoustic sensor system, which can be realized only by using the sensing units distributed along the optical fiber and preprocessing the received signal, and is simple to implement and low in cost;

2. By processing the signal, the present invention not only effectively suppresses the random noise of the system, improves the sensitivity of the system, but also has the function of enhancing or suppressing specific spatial directions and specific spatial positions, effectively expanding the distributed optical fiber acoustic sensor in sound waves Applications in the fields of communication and detection have further improved the existing sensing system's ability to monitor target signals in complex working environments, which is of great significance.

3. The present invention enables the distributed optical fiber acoustic sensor to realize real-time quantitative sound field measurement in a large range and long distance, with a large sensing aperture, easy to form a beamformer with precise spatial orientation, and can effectively suppress interference signals in the measurement environment, and improve Signal-to-interference ratio: The sensing points are densely arranged on the sensing fiber, providing a large number of redundant sensing signals for signal source enhancement, which can effectively suppress the system noise of the sensing system itself, and greatly improve the sensitivity of the sensing system .

Description of the drawings

Figure 1 is a flow chart of signal processing of the present invention;

Figure 2 is a schematic diagram of the signal processing method of the present invention;

Figure 3 is a schematic diagram of the signal processing method of the dual-parallel sensing optical fiber laying structure of the present invention;

detailed description

The present invention will be further described below with reference to the drawings and embodiments, but it is not limited thereto. According to the idea of the present invention, several implementation methods can be adopted. The following solutions are only used as explanations of the inventive idea, and the specific solutions are not limited to this.

Example 1:

The schematic diagram is shown in Figure 2. The distributed optical fiber acoustic sensor 1 including the phase-sensitive optical time domain reflectometer for coherent detection, the sensing fiber 2, the enhanced aperture 2-1, the sensing channel 2-2, and the specific spatial direction The signal sources 3-1, 3-2 and the signal source 3-3 at a specific spatial position, and the beamformer 4. The signal source can be a wideband signal or a narrowband signal. The signal processing method mainly includes the following three steps:

1) The distributed optical fiber acoustic sensor 1 of the coherent detection phase-sensitive optical time domain reflectometer emits detection light pulses to the sensing fiber 2 to quantitatively monitor the sound field sensed along the sensing fiber, And obtain the sound field distribution signal S(l,t), where l represents the one-dimensional axial space of the sensing fiber, and t is the time;

2) The phase-sensitive optical time domain reflectometer 1 preprocesses the obtained sound field signal:

First, the distributed optical fiber acoustic sensor 1 selects a section of L as the enhanced aperture 2-1 in the sensing optical fiber link, and uses the signal source spatial positioning method based on distributed optical fiber sensing in this section. The spatial position information of the received signal within the enhanced aperture 2-1 is obtained. In the enhanced aperture 2-1, the number of sensing channels 2-2 included is N, which is specifically determined by the sampling frequency f _{s of the} distributed optical fiber acoustic sensor, taking Δn as the interval.

Among them, l ₁ represents the starting position of the enhanced aperture 2-1 on the optical fiber link.

Finally, according to the spatial information of the target signal source 3-1 relative to the enhanced aperture 2-1, the time delay vector of the signal source with respect to each sensing unit is calculated

Where m=1, 2,...,M, M represents the number of signal sources.

3) The sampled signal group is input to the beamformer 4 and processed according to the delay sum method to calculate the delay compensation weight,

For other M-1 signal sources, the suppression effect is limited by the size of the enhanced aperture. If the m'th signal source needs to be enhanced, recalculate its delay compensation weight w _m' (Δτ), and calculate the enhanced signal output Y according to the above equation _m' (t) is fine.

Example 2

The principle diagram is shown in Figure 3. It includes a distributed optical fiber acoustic sensing system with an optical frequency domain reflectometer structure, a sensing fiber 2, an enhanced aperture 2-1, a sensing channel 2-2, and a signal source in a specific spatial direction. 3-1, signal source 3-2 and 3-3 beamformer 4 at a specific spatial location. The sensing optical fiber 2 is laid in a double parallel structure and is connected to the distributed optical fiber acoustic sensing system 1. The sensing optical fiber laying structure needs to meet the requirements of the signal source spatial positioning method based on distributed optical fiber sensing, which does not belong to the scope of this patent. The signal source can be considered a narrowband signal. The signal processing method mainly includes 3 steps:

1) The optical frequency domain reflectometer 1 emits detection light pulses to the sensing fiber 2, quantitatively detects the sound field sensed along the sensing fiber 2, and obtains the sound field distribution signal S(l, t), where l represents the one-dimensional axial space of the sensing fiber, and t is time;

2) The optical frequency domain reflectometer 1 preprocesses the obtained sound field signal:

First of all, the distributed optical fiber acoustic sensor 1 selects an interval of length L as the enhanced aperture 2-1 in the sensing optical fiber link. In this interval, the enhanced aperture 2-1 is specified according to the prior knowledge of the application scenario. The signal source spatial position P _m (x,y,z) of the received signal that needs to be enhanced, and the signal source spatial position P _m' (x,y,z) of the received signal that needs to be suppressed, or the received signal that needs to be enhanced Signal source space

And the signal source space of the received signal that needs to be suppressed

The number of sensing units included in the enhanced aperture is N, which is specifically determined by the spatial sampling frequency f _{s of the} distributed optical fiber acoustic sensor, taking Δn as the interval

Among them, l ₁ represents the starting position of the enhanced aperture on the optical fiber link 2_2.

According to the spatial position information or spatial direction information of the signal source to be enhanced, calculate the relative phase delay vector of each sensing unit under the spatial position or spatial direction

Where m=1, 2,...,M, M represents the number of signal sources.

3) The sampled signal group is input to the beamformer 4, and processed by the linearly constrained minimum variance beamformer (LCMV), and the covariance matrix of the sampled signal group is first calculated according to the following equation,

Where K represents the number of repetitions of the optical frequency domain reflectometer transmitting the probe light pulse to the sensing fiber.

For the m-th signal source, determine the restriction vector F _1×M =[0 0 ... 1 ... 0], where the m-th element of the restriction vector is 1, and the rest are 0. The restriction vector determines that only the m-th signal source is directional enhancement, and the other signal sources are directional suppression.

Calculate the phase compensation weight according to the following equation,

Among them, H represents conjugate transpose. For the m-th signal source, the enhanced signal is calculated according to the following equations,

Example 3

The principle diagram of the distributed optical fiber acoustic sensing system adopting the direct detection phase-sensitive optical time domain reflectometer structure is similar to that shown in Figure 2, and will not be repeated here. The signal source is a broadband signal, and its signal processing method mainly includes the following 3 steps:

1) The phase-sensitive optical time domain reflectometer 1 emits detection light pulses to the sensing fiber 2, quantitatively detects the sound field sensed along the sensing fiber 2, and obtains the sound field distribution signal S(l,t) , Where l represents the one-dimensional axial space of the sensing fiber, and t is time;

First, the distributed optical fiber acoustic sensor 1 selects a section of L as the enhanced aperture 2-1 in the sensing optical fiber link, and uses the signal source spatial positioning method based on distributed optical fiber sensing in this section. Obtain the signal source spatial position information of the received signal within the enhanced aperture. The number of sensing units included in the enhanced aperture is N, which is specifically determined by the spatial sampling frequency f _{s of the} distributed optical fiber acoustic sensor, taking Δn as the interval

Among them, l ₁ represents the starting position of the enhanced aperture on the optical fiber link 2. The sampled signal group is transformed into the frequency domain X(f), and split into P subband signals X(f _p ).

According to the spatial information of the relative enhanced aperture of the signal source, calculate the phase delay vector of each sub-band of the signal source with respect to each sensing unit

Where m=1, 2,...,M, M represents the number of signal sources.

3) The frequency-domain sampled signal group is processed by a beamformer based on linearly constrained minimum variance, and the covariance matrix corresponding to each subband is first calculated according to the following equation:

Where J represents the frequency accuracy of the fast Fourier transform performed by the direct detection optical time domain reflectometer on the time domain sampled signal group. For the m-th signal source, determine the restriction vector F _1×M =[0 0 ... 1 ... 0], where the m-th element of the restriction vector is 1, and the rest are 0. The restriction vector determines that only the m-th signal source is directional enhancement, and the other signal sources are directional suppression. Calculate the phase compensation weight corresponding to each sub-band according to the following equation,

Among them, H represents conjugate transpose. For the p-th subband of the m-th signal source, the frequency-domain enhanced signal is calculated according to the following equation,

OK.

Claims

A signal processing method based on a distributed optical fiber acoustic sensor includes a signal source, a distributed optical fiber acoustic sensor, and a sensing optical fiber. The method includes the following steps:

1) The distributed optical fiber acoustic sensor transmits detection light pulses to the sensing fiber, quantitatively monitors the sound field sensed along the sensing fiber, and obtains the sound field distribution signal S(l, t), where l Represents the one-dimensional axial space of the sensing fiber, and t is time;

2) The distributed optical fiber acoustic sensor preprocesses the obtained sound field signal:

First, select a section of length L in the sensing fiber as the enhanced aperture, and calculate the spatial position information of the signal received in the enhanced aperture;

Secondly, the number of sensing units included in the enhanced aperture is N, which is determined by the spatial sampling frequency f s of the distributed optical fiber acoustic sensor, selected at intervals of Δn,
The sound field signals collected by two sensing units are used as the sampling signal group:

Among them, l 1 represents the starting position of the enhanced aperture on the fiber link;

Finally, according to the spatial position information, calculate the time delay vector of the signal source with respect to each sensing unit in the enhanced aperture

Or phase delay vector

Where m = 1, 2, ..., M, M represents the number of signal sources;

3) Use the array signal processing method to perform signal processing on the sampled signal group and delay vector obtained in step 2).
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 1, wherein the spatial position information in the step 2) is the azimuth vector of the signal source's relative enhancement aperture or the signal source's relative enhancement aperture. Three-dimensional position coordinates.
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 1, characterized in that, in the step 2), the spatial position information is obtained using a signal source spatial positioning method or is artificially designated according to prior knowledge.
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 1, wherein the array signal processing method in the step 3) is an adaptive spatial filtering method or a delay summation method.
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 4, wherein the adaptive spatial filtering method is a minimum variance undistorted response beamformer (MVDR), linearly constrained minimum variance beamformer One of LCMV and Generalized Sidelobe Cancellation Beamformer (GSC).
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 4, wherein the delay summation method is as follows:

Calculate the delay compensation weight,

For the m-th signal source, its enhanced signal is calculated in the following way,

For other M-1 signal sources, the signal suppression effect is limited by the size of the enhanced aperture.
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 5, wherein the linearly constrained minimum variance beamformer (LCMV) method is as follows:

Calculate the covariance matrix of the sampled signal group,

Wherein K represents the number of repetitions of the distributed optical fiber acoustic sensor emitting detection light pulses to the sensing optical fiber;

For the m-th signal source, determine the restriction vector F 1×M = [0 0… 1… 0], where the m-th element of the restriction vector is 1, and the rest are 0. The restriction vector determines that only the m-th signal The source performs directional enhancement, while the other signal sources perform directional suppression,

Calculate the phase compensation weight according to the following equation,

Among them, H represents conjugate transpose,

For the m-th signal source, the enhanced signal is calculated according to the following equations,
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 1, wherein when the signal source is a broadband signal, the step 2) further comprises:

Transform the sampling signal group to the frequency domain sampling signal group X(f), and split it into P subband signals X(f p ),

According to the spatial position information of the signal source relative to the enhanced aperture, calculate the phase delay vector of each subband of the signal source with respect to each sensing unit

Or time delay vector
A signal processing method based on a distributed optical fiber acoustic sensor according to claim 8, wherein the step 3) is to use a linearly constrained minimum variance beamformer to process the frequency domain sampling signal group, specifically for:

First calculate the covariance matrix corresponding to each sub-band,

Where J represents the frequency accuracy of the fast Fourier transform of the direct detection optical time domain reflectometer on the time domain sampled signal group;

Secondly, for the m-th signal source, determine the restriction vector F 1×M = [0 0… 1… 0], where the m-th element of the restriction vector is 1, and the rest are 0. The restriction vector is determined only for the m-th The signal source performs directional enhancement, while the other signal sources perform directional suppression. The phase compensation weight corresponding to each sub-band is calculated according to the following equation,

Among them, H represents conjugate transpose,

Finally, for the p-th subband of the m-th signal source, the frequency-domain enhanced signal is calculated according to the following equation,

All the subbands are reconstructed into a complete frequency domain signal, and then through the inverse fast Fourier transform, the time domain signal of the m-th signal source after enhancement can be obtained.