CN113465528A - High-speed distributed strain measurement system and method based on optical frequency domain reflection - Google Patents

Abstract

The invention relates to a high-speed distributed strain measurement system based on optical frequency domain reflection and a measurement method thereof, wherein the measurement system comprises: tunable laser (1), 90:10 polarization maintaining beam splitter (4), computer (19), USB control line (20), modulus collection device (15), clock trigger device (22) and main interferometer (21) based on auxiliary interferometer, wherein, main interferometer (21) includes: 80:20, a polarization maintaining coupler (3), a second circulator (10), an optical mixer (13), a second balance detector (16), a third balance detector (17), a reference arm (11), a test arm (12), a tensile displacement table (14) and a sensing optical fiber (18); and the computer (19) is used for carrying out data processing on the interference signals acquired by the analog acquisition device (15) and realizing optical fiber sensing for measuring the stress of the distributed optical fiber by using long-distance optical fiber gratings, namely sensing optical fibers (18) in optical frequency domain reflection.

Description

High-speed distributed strain measurement system and method based on optical frequency domain reflection

Technical Field

The invention relates to the technical field of optical fiber sensing, in particular to a high-speed strain measurement method in optical frequency domain reflection, which is applied to optical frequency domain reflection.

Background

Compared with other traditional sensing technologies, the optical fiber sensing technology has many advantages, such as electromagnetic interference resistance, corrosion resistance, high sensitivity, high and low temperature environment resistance and the like. The method for realizing stress detection by using the optical fiber sensing technology has important application in the fields of aircraft skin monitoring, bridge structure health monitoring, perimeter security monitoring and the like. If the stress detection in the fields can realize real-time monitoring and improve the frequency of stress monitoring, the monitoring efficiency can be greatly improved, and the probability of major safety accidents is reduced. Traditional distributed optical fiber strain sensors include Brillouin Optical Time Domain Reflectometer (BOTDR) and Brillouin Optical Time Domain Analysis (BOTDA), and can realize sub-meter spatial resolution, sensing range of tens of kilometers and static/dynamic strain measurement. Fiber optic interferometer sensors, such as the MZI type and Sagnac loop type, have high sensitivity but low spatial resolution (typically tens of meters). However, the above distributed optical fiber strain sensors are generally limited to a minimum measurable strain above 10 and are not able to guarantee real-time stress monitoring with high spatial resolution, high strain resolution and high frequencies over long distances.

Optical Frequency Domain Reflectometry (OFDR) is one type of distributed Optical fiber sensing, and the OFDR technique regards rayleigh scattering in an Optical fiber as a random spatial period weak bragg grating and can be used for distributed stress and temperature sensing. In strain sensing, the OFDR adopts a Rayleigh scattering spectrum cross-correlation method, and when the spatial resolution is 1cm, the minimum measurable strain reaches +/-1 micro-strain. However, in the research of OFDR stress measurement at present, the measurement of high-frequency real-time strain with long distance and high strain resolution is still difficult to realize. In a conventional measurement device for OFDR, for example, the optical frequency domain reflection distributed sensing demodulation method based on relative phase change disclosed in patent CN201910697503.1, the measurement device adopted in the patent adjusts the polarization state of light by using a polarization controller, and adopts fifty: fifty couplers polarize the split signal. The polarization controller adopts a manual adjustment mode, so that the light intensity in two orthogonal directions is slightly deviated during polarization beam splitting, and the effect of eliminating polarization fading noise is not good enough.

Disclosure of Invention

The invention provides a high-speed distributed strain measurement system and method based on optical frequency domain reflection, which can well overcome the problems caused by polarization fading noise and can improve the data processing speed. The technical scheme is as follows:

a high-speed distributed strain measurement system based on optical frequency domain reflection comprises: tunable laser 1, 90:10 polarization maintaining beam splitter 4, computer 19, USB control line 20, analog-to-digital acquisition device 15, secondary interferometer based clock trigger device 22 and primary interferometer 21, wherein,

the auxiliary interferometer based clock trigger 22 comprises: first balanced detector 2, 50:50 a coupler 5, a delay optical fiber 6, a first Faraday rotator mirror 7, a second Faraday rotator mirror 8 and a first circulator 9; the clock trigger device 22 based on the auxiliary interferometer is used for realizing equal optical frequency interval sampling, and aims to inhibit the nonlinear scanning of a light source;

the main interferometer 21 includes: 80:20, a polarization maintaining coupler 3, a second circulator 10, an optical mixer 13, a second balance detector 16, a third balance detector 17, a reference arm 11, a test arm 12, a tensile displacement table 14 and a sensing optical fiber 18;

the output end of the USB control line 20 is connected with the input end of the tunable laser 1; the input end of the USB control line 20 is connected with the output end of the computer 19; tunable lasers 1 and 90:10, connecting ports a of the polarization maintaining optical beam splitter 4; 90: the b port of the 10 polarization maintaining beam splitter 4, namely the 10% beam splitting port, is connected with the a port of the first circulator 9; 90:10 c port of the polarization maintaining beam splitter 4, namely 90% beam splitting port, 80:20, the ports a of the polarization maintaining coupler 3 are connected; b port of the first circulator 9 and 50: port a of coupler 5 is connected with port 50; the port c of the first circulator 9 is connected with the input end of the first balanced detector 2; 50: the port b of the 50 coupler 5 is connected with the input end of the first balanced detector 2; 50: the port c of the 50 coupler 5 is connected with a first Faraday rotator mirror 7 through a delay optical fiber 6; 50: the d port of the 50 coupler 5 is connected with a second Faraday rotator mirror 8; the output end of the first balance detector 2 is connected with the input end of the analog-digital acquisition device 15; 80: the c port of the 20 polarization maintaining coupler 3, namely the 20% light splitting port, is connected with the input end a of the optical mixer 13 through the reference arm 11; 80: the d port of the 20 polarization maintaining coupler 3, namely 80% branch port, is connected with the a port of the second circulator 10 through the test arm 12; the port c of the second circulator 10 is connected with a sensing optical fiber 18; the port b of the second circulator 10 is connected with the input end b of the optical mixer 13; the output end c and the output end d of the optical mixer 13 are connected with two input ends of a second balanced detector 16; the output end e and the output end f of the optical mixer 13 are connected with two input ends of a third balanced detector 17; the output end of the second balance detector 16 is connected with the input end of the analog-digital acquisition device 15; the output end of the third balanced detector 17 is connected with the input end of the analog-digital acquisition device 15; the output end of the analog-digital acquisition device 15 is connected with the input end of the computer 19;

the computer 19 controls the tunable laser 1 to control the tuning speed, the central wavelength and the tuning start through a USB control line 20; the outgoing light of the tunable laser 1 is composed of 90:10 a port of the polarization maintaining beam splitter 4, and the ratio of 10: the ratio of 90 is from 90:10 the b port of the polarization maintaining beam splitter 4 enters 50 through the first circulator 9: 50 a-port of coupler 5, light is transmitted from 50:50 the a port of coupler 5 enters, from 50: the c and d ports of the coupler 5 of 50 are emitted, reflected by the first faraday rotator mirror 7 and the second faraday rotator mirror 8 of the two arms, respectively, and returned to the coupler 50:50 at the c, d ports of coupler 5, two beams are at 50:50 interference occurs in coupler 5, from 50:50 the b port output of coupler 5; 50: the coupler 550 enters the first balanced detector 2 from the emergent light of the port b, and the first balanced detector 2 converts the detected light signal into an interference beat frequency signal and transmits the interference beat frequency signal to the analog-digital acquisition device 15 as an external clock signal of the analog-digital acquisition device 15;

the outgoing light of the tunable laser 1 is composed of 90:10 a port of the polarization maintaining beam splitter 4, from 90:10, the c port of the optical splitter 4 enters 80:20 a port of the polarization maintaining coupler 3; passing through 80: the 20 polarization maintaining coupler 3 enters the reference arm 11 from a c port, namely a 20% optical splitting port, and enters an a port of the second circulator 10 on the test arm 12 from a d port, namely an 80% optical splitting port; light enters from the port a of the second circulator 10 and enters the sensing optical fiber 18 in the stretching region 23 from the port c of the second circulator 10, and the backscattered light of the sensing optical fiber 18 enters from the port c of the second circulator 10 and is output from the port b of the second circulator 10; the reference light output from the reference arm 11 enters an a port of the optical mixer 13 and is combined with the reference light entering a b port of the optical mixer 13 from a b port of the second circulator 10 to form beat frequency interference; the optical mixer 13 is input to the second balanced detector 16 through the port c and the port d, and the second balanced detector 16 converts the detected optical signal into an interference beat frequency signal and transmits the interference beat frequency signal to the analog-digital acquisition device 15; the optical mixer 13 is input to the third balanced detector 17 through the e port and the f port, and the second balanced detector 17 converts the detected optical signal into an interference beat frequency signal and transmits the interference beat frequency signal to the analog-digital acquisition device 15; the analog-digital acquisition device 15 transmits the acquired analog electric signal to the computer 19 under the action of an external clock signal formed by a clock trigger device 22 of the auxiliary interferometer;

the tunable laser 1 is used for providing a light source for an optical frequency domain reflection system, and the optical frequency of the tunable laser can be linearly scanned;

a first circulator 9 for preventing 50:50 the b port of coupler 5 reflects light into the laser;

50:50 coupler 5 for optical interference;

the delay optical fiber 6 is used for realizing unequal-arm beat frequency interference and can obtain optical frequency according to beat frequency and the length of the delay optical fiber;

a stretching displacement table 14 for stretching the sensing optical fiber 18 to generate controllable precise strain;

the first Faraday rotator mirror 7 and the second Faraday rotator mirror 8 are used for providing reflection for the interferometer and eliminating the polarization fading phenomenon of the interferometer;

the optical mixer 13 is used for performing polarization beam splitting on the signal, so that the light intensities of the reference light and the test light in two orthogonal directions are basically consistent during polarization beam splitting, the influence of polarization fading noise is eliminated, the combination of the reference light and the test light is realized, and beat frequency interference is formed;

and the computer 19 is used for processing the interference signals acquired by the analog acquisition device 15 and realizing optical fiber sensing for measuring the stress of the distributed optical fiber by using a long-distance optical fiber grating, namely the sensing optical fiber 18 in optical frequency domain reflection.

The sensing fiber may be a rayleigh scattering enhanced fiber or a continuous grating fiber.

The invention also provides a high-speed distributed strain measurement method based on optical frequency domain reflection, which comprises the following steps:

the first step, initialization: setting parameters of a sweep frequency mode, an initial wavelength and a termination wavelength of the tunable laser, particularly setting the sweep frequency mode as a periodic round-trip mode of forward sweep frequency from the initial wavelength to the termination wavelength and backward sweep frequency from the termination wavelength to the initial wavelength; initializing a triggering mode, an acquisition channel and data size parameters of an analog-digital acquisition device; initializing an RAM and an ROM of a CPU of a computer, dividing the RAM into an RAM1 and an RAM2, and determining the storage space size of the RAM1 and the RAM2 according to the data volume acquired by one-way frequency sweeping; initializing parameters of a GPU of a computer, and distributing memory parameters of the GPU;

secondly, the analog-digital acquisition device is triggered by the tunable laser, a clock signal generated by the auxiliary interferometer triggers to start acquiring data with equal frequency intervals, and the analog-digital acquisition device starts transmitting the data to the RAM1 of the CPU; when the storage space of the RAM1 is full, the analog-digital acquisition device transmits data to the RAM2, the CPU transmits the data in the RAM1 to the GPU for sensing information processing, and the data is used as forward frequency sweep reference data and is obtained by a tunable laser in a forward frequency sweep mode; when the RAM2 is full, the analog-digital acquisition device transmits data to the RAM1, and simultaneously the data in the RAM2 are transmitted to the GPU as reverse sweep frequency reference data, and the data are obtained by one-time reverse sweep frequency of the tunable laser; the two reference data are stored in the GPU in the whole data processing process, and calculation is carried out when the respective measured data enter the GPU in the next step; when the RAM1 is full again, the RAM1 data transmitted to the GPU is forward sweep measurement data, and when the RAM2 is full again, the RAM2 data transmitted to the GPU is reverse sweep measurement data; in the step of processing data by parallel calculation of the GPU, relative to the processing of forward sweep frequency reference data and forward sweep frequency measurement data, the processing of reverse sweep frequency reference data and reverse sweep frequency measurement data is the same as the forward processing flow except that negation is required for solving the maximum position of a cross correlation peak;

thirdly, dividing the forward sweep frequency reference data and the forward sweep frequency measurement data into k sections according to the sweep frequency length of the tunable laser in the GPU, wherein the processing flow of each section of data is the same;

step four, the data type is forced to convert the kernel function: for the operation of the subsequent kernel function, performing forced conversion of data types from short to float on reference data and measurement data which enter the GPU after segmentation;

fifthly, fast forward Fourier transform library function: performing forward fast Fourier transform on each section of reference data and measurement data after data conversion, and converting the reference data and the measurement data into complex signals corresponding to each position along the optical fiber testing distance;

sixthly, piecewise zero padding kernel function: taking the number of segmented window points as N, taking the zero-filling point of each segment of data as M, and dividing each segment of reference data and measurement data subjected to fast forward Fourier transform into data with each M + N points as a group in the zero-filling process;

step seven, segmenting a fast inverse Fourier transform library function: carrying out fast inverse Fourier transform on each M + N points of the reference data and the measurement data after the segmentation zero padding, and converting the reference data and the measurement data from a distance domain to complex signals on a frequency domain;

and eighthly, taking an amplitude kernel function: amplitude values of the reference data and the measured data processed in the seventh step are obtained;

and ninthly, reducing a mean value removing kernel function to remove direct current components in the reference data and the measurement data: executing the kernel function in the GPU requires thread blocks and threads, each thread corresponds to one piece of data, each thread block is composed of a plurality of threads, and through the seventh step, zero padding is carried out on each group of M + N points of two groups of data in advance, so that each section becomes an integral multiple of the number of the threads in the block; then dividing each small segment of data after zero padding into a plurality of thread blocks for specification summation; obtaining a result in each thread block, summing a plurality of results in each small segment of data in a circulating mode to obtain the respective sum of each small segment, and subtracting the average from the data of each small segment to obtain the reference data and the measured data of which the direct current items are removed;

step ten, cross-correlation kernel function, which utilizes convolution theorem to transform cross-correlation algorithm into multiplication of distance domain for calculation: processing every M + N points of the reference data and the measurement data processed in the ninth step, firstly, turning every M + N points of the reference data and filling zero by one time; secondly, zero filling is carried out on every M + N test data by one time; then, simultaneously carrying out forward fast Fourier transform on the two groups of data, multiplying the two groups of data, and carrying out reverse fast Fourier transform on a product result;

the eleventh step: peak finding kernel function: performing cross-correlation operation on a group of M + N points corresponding to the reference data and the measured data in the tenth step, calculating to finish 2 points (M + N), searching the position of the peak value of each 2 points (M + N), determining the optical frequency domain movement result, transmitting the data result back to a ROM memory of a CPU by a GPU for storage, and obtaining the strain information on the primary optical fiber;

for each section of segmented data, processing the segmented forward sweep frequency reference data and forward sweep frequency measurement data according to the modes from the third step to the tenth step, so that k times of strain information can be obtained from data stored in one RAM; the reverse sweep frequency reference data and the reverse sweep frequency measurement data are also divided into k sections of data to be processed according to the steps, and the inversion operation is only carried out on the maximum value position of the peak when the peak is searched in the eleventh step.

In the process of processing the sensing data by the GPU kernel function, except that partial statements of the specification mean value-removing kernel function and the peak-finding kernel function need to be subjected to serial operation, the rest kernel functions and library functions are subjected to extremely high-speed parallel computation; when the data acquisition module transmits data to one RAM partition of the CPU, the GPU processes the data of the other RAM partition at the same time, so that the whole strain measurement process is real-time and high-speed, and the strain data are processed while the data are acquired. The technical scheme provided by the invention has the beneficial effects that:

1. measurement of 20hz sinusoidal strain with resolution of 20cm for real time measurement of 200m fiber is achieved.

2. Distributed fiber strain measurements with a minimum measured strain peak to peak of 2.5 mu epsilon and an R value of 0.9972 high linearity system were achieved.

Drawings

FIG. 1 is a high-speed distributed strain measurement system based on optical frequency domain reflection according to the present invention;

FIG. 2 is a schematic illustration of a calibration curve;

FIG. 3 is a schematic diagram of a stretching displacement table stretching a full grating fiber;

FIG. 4 is a two-dimensional strain diagram and a spectrogram at different frequencies;

FIG. 5 is a two-dimensional strain diagram and a spectrogram under a small strain;

fig. 6 is a flow chart of a high-speed strain measurement system based on optical frequency domain reflection distributed sensing.

In the drawings, the components represented by the respective reference numerals are listed below:

1: a tunable laser; 2: a first balanced detector;

3: 80:20 polarization maintaining coupler; 4: 90:10 polarization maintaining beam splitter;

5: a 50:50 coupler; 6: a delay optical fiber;

7: a first Faraday rotator mirror; 8: a second Faraday rotator mirror;

9: a first circulator; 10: a second circulator;

11: a reference arm; 12: a test arm;

13: a light mixer; 14: stretching the displacement table;

15: a modulus acquisition device; 16: a second balanced detector;

17: a third balanced detector; 18: a sensing optical fiber;

19: a computer; 20: a USB control line.

Detailed Description

The first embodiment is as follows:

this example includes a high-speed distributed strain measurement system based on optical frequency domain reflection.

The high-speed distributed strain measurement system based on optical frequency domain reflection comprises: the system comprises a tunable laser 1, a 90:10 polarization maintaining beam splitter 4, a computer 19, a USB control line 20, an analog-digital acquisition device 15, a clock trigger device 22 based on an auxiliary interferometer and a main interferometer 21.

The clock trigger device 22 based on the auxiliary interferometer includes: a first balanced detector 2, a 50:50 coupler 5, a delay fiber 6, a first faraday rotator mirror 7, a second faraday rotator mirror 8 and a first circulator 9. The auxiliary interferometer based clock trigger 22 is used to achieve equal optical frequency spacing sampling with the aim of suppressing non-linear scanning of the light source.

Wherein the main interferometer 21 includes: 80:20 polarization maintaining coupler 3, second circulator 10, optical mixer 13, second balance detector 16, third balance detector 17, reference arm 11, test arm 12, tension displacement table 14 and sensing fiber 18, which is Rayleigh scattering enhanced fiber or continuous grating fiber. The main interferometer 21 is the core of a distributed fiber optic sensing device that reflects in the optical frequency domain, which is a modified mach zehnder interferometer.

The output end of the USB control line 20 is connected with the input end of the tunable laser 1; the input end of the USB control line 20 is connected with the output end of the computer 19; the tunable laser 1 is connected with the port a of the 90:10 polarization-maintaining beam splitter 4; the port b of the 90:10 polarization-maintaining beam splitter 4, namely a 10% beam splitting port, is connected with the port a of the first circulator 9; the port c of the 90:10 polarization-maintaining optical beam splitter 4, namely the 90% optical splitting port, is connected with the port a of the 80:20 polarization-maintaining coupler 3; the b port of the first circulator 9 is connected with the a port of the 50:50 coupler 5; the port c of the first circulator 9 is connected with the input end of the first balanced detector 2; the port b of the 50:50 coupler 5 is connected with the input end of the first balanced detector 2; the port c of the 50:50 coupler 5 is connected with a first Faraday rotator mirror 7 through a delay optical fiber 6; the d port of the 50:50 coupler 5 is connected with the second Faraday rotator mirror 8; the output end of the first balance detector 2 is connected with the input end of the analog-digital acquisition device 15; the c port of the 80:20 polarization maintaining coupler 3, namely the 20% light splitting port, is connected with the input end a of the optical mixer 13 through the reference arm 11; the d port of the 80:20 polarization-maintaining coupler 3, namely the 80% branch port, is connected with the a port of the second circulator 10 through the test arm 12; the port c of the second circulator 10 is connected with a sensing optical fiber 18; the port b of the second circulator 10 is connected with the input end b of the optical mixer 13; the output end c and the output end d of the optical mixer 13 are connected with two input ends of a second balanced detector 16; the output end e and the output end f of the optical mixer 13 are connected with two input ends of a third balanced detector 17; the output end of the second balance detector 16 is connected with the input end of the analog-digital acquisition device 15; the output end of the third balanced detector 17 is connected with the input end of the analog-digital acquisition device 15; the output end of the analog-digital acquisition device 15 is connected with the input end of the computer 19.

When the device works, the computer 19 controls the tunable laser 1 to control the tuning speed, the central wavelength, the tuning start and the like through the USB control line 20; outgoing light of the tunable laser 1 enters from a port a of a 90:10 polarization-maintaining beam splitter 4, enters from a port b of the 90:10 polarization-maintaining beam splitter 4 into a port a of a 50:50 coupler 5 through a first circulator 9 in a ratio of 10:90, enters from the port a of the 50:50 coupler 5, exits from ports c and d of the 50:50 coupler 5, is reflected by a first Faraday rotator 7 and a second Faraday rotator 8 of two arms respectively, returns to ports c and d of the 50:50 coupler 5, interferes in the 50:50 coupler 5, and is output from the port b of the 50:50 coupler 5; 50: the coupler 550 enters the first balanced detector 2 from the emergent light of the b port, and the first balanced detector 2 converts the detected light signal into an interference beat frequency signal and transmits the interference beat frequency signal to the analog-digital acquisition device 15 as an external clock signal of the analog-digital acquisition device 15.

Emergent light of the tunable laser 1 enters from a port a of a 90:10 polarization-maintaining optical beam splitter 4 and enters from a port c, namely a 90% optical splitting port, of the 90:10 optical beam splitter 4 into a port a of an 80:20 polarization-maintaining coupler 3; the polarization maintaining coupler 3 enters the reference arm 11 from a port c, namely a 20% optical splitting port, and enters a port a of the second circulator 10 on the test arm 12 from a port d, namely an 80% optical splitting port through an 80:20 polarization maintaining coupler; light enters from the port a of the second circulator 10 and enters the sensing optical fiber 18 in the stretching region 23 from the port c of the second circulator 10, and the backscattered light of the sensing optical fiber 18 enters from the port c of the second circulator 10 and is output from the port b of the second circulator 10; the reference light output from the reference arm 11 enters an a port of the optical mixer 13 and is combined with the reference light entering a b port of the optical mixer 13 from a b port of the second circulator 10 to form beat frequency interference; the optical mixer 13 is input to the second balanced detector 16 through the port c and the port d, and the second balanced detector 16 converts the detected optical signal into an interference beat frequency signal and transmits the interference beat frequency signal to the analog-digital acquisition device 15; the optical mixer 13 is input to the third balanced detector 17 through the e port and the f port, and the second balanced detector 17 converts the detected optical signal into an interference beat frequency signal and transmits the interference beat frequency signal to the analog-digital acquisition device 15; the analog-to-digital acquisition device 15 transmits the acquired analog electrical signal to the computer 19 under the action of an external clock signal formed by a clock trigger device 22 of the auxiliary interferometer.

The USB control line 20 is used by the computer 19 to control the tunable laser 1 through it.

The tunable laser 1 provides a light source for an optical frequency domain reflection system, the optical frequency of which can be scanned linearly.

The first circulator 9 prevents the b-port reflected light of the 50:50 coupler 5 in the auxiliary interferometer from entering the laser.

The 50:50 coupler 5 is used for optical interference.

The delay fiber 6 is used to realize beat frequency interference of an unequal arm, and can obtain an optical frequency according to the beat frequency and the length of the delay fiber.

The tension displacement stage 14 is used to tension the sensing fiber 18 to produce a controlled fine strain.

The first Faraday rotator mirror 7 and the second Faraday rotator mirror 8 are used for providing reflection for the interferometer and eliminating the polarization fading phenomenon of the interferometer.

The optical mixer 13 performs polarization beam splitting on the signal, so that the light intensities of the reference light and the test light in two orthogonal directions are substantially consistent during polarization beam splitting, the influence of polarization fading noise is eliminated, the combination of the reference light and the test light is realized, and beat frequency interference is formed.

The computer 19: and the interference signals acquired by the analog acquisition device 15 are subjected to data processing, so that optical fiber sensing for measuring the stress of the distributed optical fiber by using the long-distance optical fiber grating in optical frequency domain reflection is realized.

Example two

The embodiment provides a high-speed distributed strain measurement method based on optical frequency domain reflection, which comprises the following steps:

the first step, initialization: the method comprises the steps of setting parameters such as a frequency sweep mode, an initial wavelength, a termination wavelength and the like of the tunable laser, and particularly setting the frequency sweep mode of the tunable laser to be a periodic round-trip mode from the initial wavelength to the termination wavelength (forward frequency sweep) and from the termination wavelength to the initial wavelength (reverse frequency sweep) in order to improve the frequency sweep efficiency of the tunable laser, so that the time for rescanning the laser from the termination wavelength to the initial wavelength in one-way frequency sweep can be saved. Initializing parameters such as a triggering mode, an acquisition channel, data size and the like of the analog-digital acquisition device. A RAM (random access memory) and a ROM (read only memory) of a computer CPU are initialized, and the RAM is divided into a RAM1 and a RAM 2. Initializing the parameters of the GPU of the computer, and distributing the parameters such as the memory of the GPU.

Secondly, after the initialization of the CPU of the computer is finished, the analog-digital acquisition device is triggered by the tunable laser, the clock signal generated by the auxiliary interferometer triggers the acquisition of data with equal frequency intervals, and then the analog-digital acquisition device starts to transmit the data to the RAM1 of the CPU. When the storage space of the RAM1 is full, the analog-digital acquisition device transmits data to the RAM2, at this time, the CPU transmits the data in the RAM1 to the GPU for sensing information processing, and the data is obtained by the tunable laser in one forward frequency sweep as forward frequency sweep reference data. When the RAM2 is full, the analog-digital acquisition device transmits data to the RAM1, and simultaneously the data in the RAM2 is transmitted to the GPU as inverse sweep reference data, which is obtained by inverse sweep of the tunable laser once. The two reference data are stored in the GPU in the whole data processing process, and the calculation is carried out until the next time the respective measured data enter the GPU. When RAM1 is full again, the RAM1 data transferred to the GPU is forward sweep measurement data, and when RAM2 is full again, the RAM2 data transferred to the GPU is reverse sweep measurement data. In the following steps of processing data by the GPU in parallel computing, we take the processing of the forward sweep reference data and the forward sweep measurement data as an example, i.e. the RAM transmits the solid arrow part in the processing flow chart of fig. 6. The processing of the reverse frequency sweep reference data and the reverse frequency sweep measurement data, namely the dotted arrow part in RAM transmission, has the steps which are the same as the forward processing flow except that the negation is required to be carried out when the maximum value position of the cross correlation peak is obtained in the tenth step.

And thirdly, processing the sensing information data obtained in the second step by different kernel functions and library functions in a GPU, and dividing two groups of data (forward sweep frequency reference data and forward sweep frequency measurement data) into six sections according to the sweep frequency length of the tunable laser for processing so as to improve the utilization efficiency of the tunable laser and improve the frequency of strain measurement. The processing flow of the six pieces of data is completely the same, and for convenience of description of the processing flow, the first piece of forward sweep reference data and the first piece of forward sweep measurement data are referred to as reference data and measurement data.

Step four, the data type is forced to convert the kernel function: and for the operation of the subsequent kernel function, performing forced conversion of the data type from short to float on the reference data and the measurement data which enter the GPU after segmentation.

Fifthly, fast forward Fourier transform library function: and performing forward fast Fourier transform on the reference data and the measured data processed in the second step, and converting the reference data and the measured data into complex signals corresponding to all positions along the optical fiber testing distance.

Sixthly, piecewise zero padding kernel function: the number of segmented window points is N, the zero filling point of each segment of data is M, and the zero filling process actually carries out interpolation on frequency domain signals, so that higher strain resolution can be realized. Thus, the reference data and the measurement data after the fast forward fourier transform are respectively divided into a group of data of every M + N points.

Step seven, segmenting a fast inverse Fourier transform library function: and performing fast inverse Fourier transform on each M + N points of the reference data and the measured data after the step of zero padding, and converting the reference data and the measured data into complex signals on a frequency domain from a distance domain.

And eighthly, taking an amplitude kernel function: and amplitude values are taken for the reference data and the measured data processed in the seventh step.

And ninthly, reducing a mean value removing kernel function: this step will remove the dc component of the reference data and the measured data. The execution of kernel functions in the GPU requires thread blocks and threads, each thread corresponding to a piece of data, a thread block being composed of multiple threads. But the thread blocks cannot be synchronized (i.e. mutual access of data cannot be realized), so we generally use the thread blocks as units to sum up the data. Therefore, we need to perform zero padding in advance for each set of M + N points for each of the two sets of data, so that each segment becomes an integer multiple of the number of threads in the block. Then, we divide each small segment of data after zero padding into multiple thread blocks for specification and summation. And each thread block will get a result, and we sum up several results in each small segment of data in a round-robin manner to get the respective sum of each small segment. Subtracting the average from the data of each segment to obtain the reference data and the measured data of which the direct current terms are removed.

Tenth step, cross-correlation kernel function: we calculate by using the convolution theorem to transform the cross-correlation algorithm into a multiplication in the distance domain. Here, we process every M + N points of the reference data and the measurement data after the ninth step, and first, zero-filling every M + N flips of the reference data. Second, zero is doubled for every M + N test data. And then, simultaneously carrying out forward fast Fourier transform on the two groups of data, multiplying the two groups of data, and carrying out reverse fast Fourier transform on a product result.

The eleventh step: peak finding kernel function: and performing cross-correlation operation on a group of M + N points corresponding to the reference data and the measured data in the tenth step, and generating 2 points (M + N) after the calculation is finished. And searching the position of each 2-point (M + N) peak value to determine the optical frequency domain movement result. And transmitting the data result back to a ROM memory of the CPU by the GPU for storage to obtain the strain information on the primary optical fiber.

The third step to the tenth step are to process the first section of the forward sweep reference data and the forward sweep measurement data, and the forward sweep data of the other five sections are also processed according to the steps, so that the data stored in the RAM can obtain the six-time strain information. The reverse sweep frequency reference data and the reverse sweep frequency measurement data are also divided into six sections of data to be processed according to the steps, and the inversion operation is only carried out on the maximum value position of the peak when the peak is searched in the eleventh step.

In the process of processing the sensing data by the GPU kernel function, except that partial statements of the specification mean value removing kernel function and the peak searching kernel function need to be subjected to serial operation, the rest kernel functions and library functions are subjected to parallel computation with extremely high speed. When the data acquisition module transmits data to one RAM partition of the CPU, the GPU processes the data of the other RAM partition at the same time, so that the whole strain measurement process is real-time and high-speed, and the strain data are processed while the data are acquired.

This patent proposes a high-speed distributed strain measurement system and method based on optical frequency domain reflection. In the device, a structure of a polarization-maintaining fiber coupler and an optical mixer is adopted, so that the defects of using a polarization controller and a 50:50 coupler can be overcome, and polarization diversity detection of a main interferometer is realized. When in polarization beam splitting, the light intensity in two orthogonal directions has small error, and the problem caused by polarization fading noise can be well solved. In order to further improve the data processing speed, parallel operation of data is achieved through a kernel function and a library function in a GPU, and in addition, data transmission and data processing are parallel through a transmission program of a collection card, so that the data speed is greatly improved, and the measuring speed reaches 60 Hz. Based on this we achieved a measurement of 20Hz vibration frequency sinusoidal strain with a spatial resolution of 20cm for a 200m fiber in real time, with a minimum measured strain peak to peak of 2.5 mu epsilon.

The feasibility of the high-speed strain frequency measurement of the sensing system is verified in conjunction with specific tests, see fig. 4, described in detail below:

the verification experiment of the embodiment of the invention adopts the sensing optical fiber 18 as a continuous grating optical fiber, the long-distance optical fiber grating is 200m in length and consists of 20000 sections, each section is 10mm in length, the grating length is 9mm, and the central wavelength is 1550 nm.

To verify the linear relationship of the system measurements, we first calibrated the system. We measured the optical frequency domain shift at static strain 2.5. mu. epsilon., 7.5. mu. epsilon., 12.5. mu. epsilon., 17.5. mu. epsilon., 22.5. mu. epsilon., 27.5. mu. epsilon. The relationship is shown in FIG. 2: from the figure, it can be seen that the system has high linearity after fitting, the R value reaches 0.9972, and the optical frequency domain shift and the stress have a fixed coefficient relationship.

One end of the sensing optical fiber 18 with the tail end of 100cm is fixed, the other end of the sensing optical fiber with the distance of 40cm is stuck on a stretching displacement table at an interval of 20cm, and then the rest 40cm of optical fiber is fixed on the stretching displacement table, as shown in figure 3. We applied a sinusoidal strain to the tensile displacement stage with a peak to peak value of 27.5 μ s, based on 15 μ s. The frequencies are 12hz, 16hz, 20hz, while the theoretical measured maximum value of the whole system is 30 hz.

In the experiment, an OFDR system is used, the starting frequency is 1551nm, the terminating frequency is 1601nm, the sweep rate is 500nm/s, the number of sampling points is 15M, and the length of an additional interferometer optical fiber is 500M.

As can be seen from a, c, and d of fig. 4, since only two positions are provided with strain changes in the experiment, there are two color stripes with color changes, and the color changes of the color stripes represent the changes in strain. And because the strain length at each position is 40cm, each colored ribbon has two points. Since the two strains are 20cm apart, the color bands are 1 strain-free point apart. As viewed on the time axis, the color changes more and more densely as the frequency increases.

The result of the fourier transform is shown in fig. 4 b, e, and f. Positions a and B represent the spectrogram of two different positions. As the frequency increases, the peaks of the spectrogram change continuously, 12hz, 16hz, and 20hz respectively, which is exactly the same as the frequency we set for, and the spectrograms at the two positions a and B are substantially the same. Therefore, we determined that the system has the ability to measure strain at a frequency of 20 hz.

Example III

To measure the ultimate strain resolution of the system, we measured sinusoidal strains of 20hz, 12.5 μ ε baseline stress, 3.75 μ ε peak to peak, 2.5 μ ε peak, 1.25 μ ε peak to peak, and the results are shown in FIG. 5.

From a, c and e in fig. 5, as the strain is reduced, the strain change can be seen to be reduced, and the system reaches the limit gradually. At 3.75. mu. epsilon, the color alternates and a change in strain is visible. At 2.5 mu epsilon, the color is similar making the change less noticeable as the stress becomes smaller, but the same frequency change remains. At 1.25. mu. epsilon. the color was uniform and no strain could be measured.

Furthermore, from the spectrograms of b, d and f in fig. 5, we can know that the peak of the spectrum is continuously reduced along with the continuous reduction of the strain, and we can not observe the peak of the spectrum of 20hz when the strain is 1.25 μ epsilon. Thus we can obtain that the system can measure strain with a peak to peak strain of 2.5 μ ε at a minimum frequency of 20 hz.

In summary, the system is based on GPU parallel operation, and realizes measurement of 20hz sinusoidal strain with resolution of 20cm by program optimization, wherein the measurement is carried out on 200m optical fiber in real time, and the minimum measurement strain peak value is 2.5 mu epsilon. According to the fitting situation, the linearity degree of the obtained system is high, and the R value reaches 0.9972.

In the embodiment of the present invention, except for the specific description of the model of each device, the model of other devices is not limited, as long as the device can perform the above functions. Those skilled in the art will appreciate that the drawings are only schematic illustrations of preferred embodiments, and the above-described embodiments of the present invention are merely provided for description and do not represent the merits of the embodiments. The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (3)

1. A high-speed distributed strain measurement system based on optical frequency domain reflection comprises: tunable laser (1), 90: the device comprises a polarization-maintaining optical beam splitter (4), a computer (19), a USB control line (20), an analog-digital acquisition device (15), a clock trigger device (22) based on an auxiliary interferometer and a main interferometer (21). Wherein the content of the first and second substances,

an auxiliary interferometer based clock trigger (22) comprising: first balanced detector (2), 50:50 coupler (5), delay fiber (6), first Faraday rotator mirror (7), second Faraday rotator mirror (8) and first circulator (9); the clock trigger device (22) based on the auxiliary interferometer is used for realizing equal optical frequency interval sampling, and aims to inhibit the nonlinear scanning of a light source;

the main interferometer (21) comprises: 80:20, a polarization maintaining coupler (3), a second circulator (10), an optical mixer (13), a second balance detector (16), a third balance detector (17), a reference arm (11), a test arm (12), a tensile displacement table (14) and a sensing optical fiber (18);

the output end of the USB control line (20) is connected with the input end of the tunable laser (1); the input end of the USB control line (20) is connected with the output end of the computer (19); tunable laser (1) and 90:10 the ports a of the polarization maintaining optical beam splitter (4) are connected; 90: the b port, namely a 10% light splitting port, of the 10 polarization-maintaining beam splitter (4) is connected with the a port of the first circulator (9); 90:10 c port of polarization maintaining beam splitter (4), namely 90% beam splitting port, 80:20, the ports a of the polarization maintaining coupler (3) are connected; b port of the first circulator (9) and 50: the ports a of the 50 couplers (5) are connected; the port c of the first circulator (9) is connected with the input end of the first balanced detector (2); 50: the port b of the 50 coupler (5) is connected with the input end of the first balanced detector (2); 50: the port c of the 50 coupler (5) is connected with a first Faraday rotator mirror (7) through a delay optical fiber (6); 50: the d port of the 50 coupler (5) is connected with a second Faraday rotator mirror (8); the output end of the first balance detector (2) is connected with the input end of the analog-digital acquisition device (15); 80: the c port of the 20 polarization maintaining coupler (3), namely the 20% light splitting port, is connected with the input end a of the optical mixer (13) through a reference arm (11); 80: the d port, namely 80% of branch port, of the 20 polarization-maintaining coupler (3) is connected with the a port of the second circulator (10) through a test arm (12); the port c of the second circulator (10) is connected with a sensing optical fiber (18); the port b of the second circulator (10) is connected with the input end b of the optical mixer (13); the output end c and the output end d of the light mixer (13) are connected with two input ends of a second balanced detector (16); the output end e and the output end f of the light mixer (13) are connected with two input ends of a third balanced detector (17); the output end of the second balance detector (16) is connected with the input end of the analog-digital acquisition device (15); the output end of the third balanced detector (17) is connected with the input end of the analog-digital acquisition device (15); the output end of the analog-digital acquisition device (15) is connected with the input end of the computer (19);

the computer (19) controls the tunable laser (1) to control the tuning speed, the central wavelength and the tuning start through a USB control line (20); the emergent light of the tunable laser (1) is composed of a 90:10 a port of the polarization-maintaining beam splitter (4), and the ratio of 10: the ratio of 90 is from 90:10 b port of polarization maintaining beam splitter (4) enters 50 through first circulator (9): 50 a-port of coupler (5), light is transmitted from 50: the a port of the 50 coupler (5) enters from the 50: the c and d ports of the 50 coupler (5) are emitted, reflected by the first Faraday rotator mirror (7) and the second Faraday rotator mirror (8) of the two arms, and returned to the 50: at the c and d ports of the 50 coupler (5), two beams of light are reflected at the 50:50 interference occurs in the coupler (5), from 50:50 the b port output of the coupler (5); 50: the coupler (5) 50 enters a first balanced detector (2) from emergent light of a port b, the first balanced detector (2) converts a detected light signal into an interference beat frequency signal and transmits the interference beat frequency signal to an analog-digital acquisition device (15) to serve as an external clock signal of the analog-digital acquisition device (15);

the emergent light of the tunable laser (1) is composed of a 90:10 a port of the polarization-maintaining beam splitter (4), from 90:10, the c port of the optical beam splitter (4), namely a 90% light splitting port enters the optical beam splitter 80:20 a port of the polarization maintaining coupler (3); passing through 80: the 20 polarization-maintaining coupler (3) enters the reference arm (11) from a c port, namely a 20% optical splitting port, and enters an a port of a second circulator (10) on the test arm (12) from a d port 80% optical splitting port; light enters from the a port of the second circulator (10) and enters the sensing optical fiber (18) in the stretching area (23) from the c port of the second circulator (10), and the back scattering light of the sensing optical fiber (18) enters from the c port of the second circulator (10) and is output from the b port of the second circulator (10); the reference light output from the reference arm (11) enters an a port of the optical mixer (13) and the reference light entering a b port of the optical mixer (13) from a b port of the second circulator (10) are combined to form beat frequency interference; the light mixer (13) is input into a second balance detector (16) through a port c and a port d, and the second balance detector (16) converts the detected light signals into interference beat frequency signals and transmits the interference beat frequency signals to an analog-digital acquisition device (15); the light mixer (13) is input into a third balanced detector (17) through an e port and an f port, and the second balanced detector (17) converts the detected light signals into interference beat frequency signals and transmits the interference beat frequency signals to an analog-digital acquisition device (15); the analog-digital acquisition device (15) transmits the acquired analog electric signal to the computer (19) under the action of an external clock signal formed by a clock trigger device (22) of the auxiliary interferometer;

a tunable laser (1) that provides a light source for an optical frequency domain reflection system, the optical frequency of which can be linearly scanned;

a first circulator (9) for preventing 50:50 the b port of the coupler (5) reflects light into the laser;

50:50 coupler (5) for optical interference;

the delay optical fiber (6) is used for realizing unequal-arm beat frequency interference and can obtain optical frequency according to beat frequency and the length of the delay optical fiber;

the stretching displacement table (14) is used for stretching the sensing optical fiber (18) to generate controllable precise strain;

the first Faraday rotator mirror (7) and the second Faraday rotator mirror (8) are used for providing reflection for the interferometer and eliminating the polarization fading phenomenon of the interferometer;

the optical mixer (13) is used for carrying out polarization beam splitting on the signal, so that the light intensities of the reference light and the test light in two orthogonal directions are basically consistent during polarization beam splitting, the influence of polarization fading noise is eliminated, the combination of the reference light and the test light is realized, and beat frequency interference is formed;

and the computer (19) is used for carrying out data processing on the interference signals acquired by the analog acquisition device (15) and realizing optical fiber sensing for measuring the stress of the distributed optical fiber by using long-distance optical fiber gratings, namely sensing optical fibers (18) in optical frequency domain reflection.

2. The measurement system of claim 1, wherein the sensing fiber is a rayleigh scattering enhanced fiber or a continuous grating fiber.

3. A high-speed distributed strain measurement method based on optical frequency domain reflection comprises the following steps:

the first step, initialization: setting parameters of a sweep frequency mode, an initial wavelength and a termination wavelength of the tunable laser, particularly setting the sweep frequency mode as a periodic round-trip mode of forward sweep frequency from the initial wavelength to the termination wavelength and backward sweep frequency from the termination wavelength to the initial wavelength; initializing a triggering mode, an acquisition channel and data size parameters of an analog-digital acquisition device; initializing an RAM and an ROM of a CPU of a computer, dividing the RAM into an RAM1 and an RAM2, and determining the storage space size of the RAM1 and the RAM2 according to the data volume acquired by one-way frequency sweeping; initializing parameters of a GPU of a computer, and distributing memory parameters of the GPU;

secondly, the analog-digital acquisition device is triggered by the tunable laser, a clock signal generated by the auxiliary interferometer triggers to start acquiring data with equal frequency intervals, and the analog-digital acquisition device starts transmitting the data to the RAM1 of the CPU; when the storage space of the RAM1 is full, the analog-digital acquisition device transmits data to the RAM2, the CPU transmits the data in the RAM1 to the GPU for sensing information processing, and the data is used as forward frequency sweep reference data and is obtained by a tunable laser in a forward frequency sweep mode; when the RAM2 is full, the analog-digital acquisition device transmits data to the RAM1, and simultaneously the data in the RAM2 are transmitted to the GPU as reverse sweep frequency reference data, and the data are obtained by one-time reverse sweep frequency of the tunable laser; the two reference data are stored in the GPU in the whole data processing process, and calculation is carried out when the respective measured data enter the GPU in the next step; when the RAM1 is full again, the RAM1 data transmitted to the GPU is forward sweep measurement data, and when the RAM2 is full again, the RAM2 data transmitted to the GPU is reverse sweep measurement data; in the step of processing data by parallel calculation of the GPU, relative to the processing of forward sweep frequency reference data and forward sweep frequency measurement data, the processing of reverse sweep frequency reference data and reverse sweep frequency measurement data is the same as the forward processing flow except that negation is required for solving the maximum position of a cross correlation peak;

thirdly, dividing the forward sweep frequency reference data and the forward sweep frequency measurement data into k sections according to the sweep frequency length of the tunable laser in the GPU, wherein the processing flow of each section of data is the same;

step four, the data type is forced to convert the kernel function: for the operation of the subsequent kernel function, performing forced conversion of data types from short to float on reference data and measurement data which enter the GPU after segmentation;

fifthly, fast forward Fourier transform library function: performing forward fast Fourier transform on each section of reference data and measurement data after data conversion, and converting the reference data and the measurement data into complex signals corresponding to each position along the optical fiber testing distance;

sixthly, piecewise zero padding kernel function: taking the number of segmented window points as N, taking the zero-filling point of each segment of data as M, and dividing each segment of reference data and measurement data subjected to fast forward Fourier transform into data with each M + N points as a group in the zero-filling process;

step seven, segmenting a fast inverse Fourier transform library function: carrying out fast inverse Fourier transform on each M + N points of the reference data and the measurement data after the segmentation zero padding, and converting the reference data and the measurement data from a distance domain to complex signals on a frequency domain;

and eighthly, taking an amplitude kernel function: amplitude values of the reference data and the measured data processed in the seventh step are obtained;

and ninthly, reducing a mean value removing kernel function to remove direct current components in the reference data and the measurement data: executing the kernel function in the GPU requires thread blocks and threads, each thread corresponds to one piece of data, each thread block is composed of a plurality of threads, and through the seventh step, zero padding is carried out on each group of M + N points of two groups of data in advance, so that each section becomes an integral multiple of the number of the threads in the block; then dividing each small segment of data after zero padding into a plurality of thread blocks for specification summation; obtaining a result in each thread block, summing a plurality of results in each small segment of data in a circulating mode to obtain the respective sum of each small segment, and subtracting the average from the data of each small segment to obtain the reference data and the measured data of which the direct current items are removed;

step ten, cross-correlation kernel function, which utilizes convolution theorem to transform cross-correlation algorithm into multiplication of distance domain for calculation: processing every M + N points of the reference data and the measurement data processed in the ninth step, firstly, turning every M + N points of the reference data and filling zero by one time; secondly, zero filling is carried out on every M + N test data by one time; then, simultaneously carrying out forward fast Fourier transform on the two groups of data, multiplying the two groups of data, and carrying out reverse fast Fourier transform on a product result;

the eleventh step: peak finding kernel function: performing cross-correlation operation on a group of M + N points corresponding to the reference data and the measured data in the tenth step, calculating to finish 2 points (M + N), searching the position of the peak value of each 2 points (M + N), determining the optical frequency domain movement result, transmitting the data result back to a ROM memory of a CPU by a GPU for storage, and obtaining the strain information on the primary optical fiber;

for each section of segmented data, processing the segmented forward sweep frequency reference data and forward sweep frequency measurement data according to the modes from the third step to the tenth step, so that k times of strain information can be obtained from data stored in one RAM; the reverse sweep frequency reference data and the reverse sweep frequency measurement data are also divided into k sections of data to be processed according to the steps, and the inversion operation is only carried out on the maximum value position of the peak when the peak is searched in the eleventh step.

Images