Optical fiber polarization mode dispersion distributed measurement method and system

Abstract

The invention provides an optical fiber polarization mode dispersion distributed measurement method and system, and the method comprises the steps: enabling pulse polarized light to enter a to-be-measured optical fiber with unknown parameters for transmission, and enabling the polarization state of the pulse polarized light to randomly change at different positions of the to-be-measured optical fiber when the pulse polarized light is transmitted in the to-be-measured optical fiber due to a birefringence effect; detecting the polarization state of Rayleigh scattering light of the pulse polarized light, and constructing a polarized light time domain reflection curve; effective features of the polarized light time domain reflection curve are extracted and input into a pre-trained machine learning model, a polarization mode dispersion prediction result of the to-be-measured optical fiber is obtained, and distributed polarization mode dispersion measurement is achieved; compared with a traditional distributed polarization mode dispersion measurement scheme, the method only needs to emit the detection pulse light once, is low in cost and high in precision, does not need to change the optical signal polarization state of the receiving end for multiple times, is simple in system and convenient to install, and effectively improves the measurement efficiency of distributed polarization mode dispersion.

Description

Technical Field

[0001] The present invention relates to the technical field of optical fiber sensing, and more specifically, to a method and system for distributed measurement of optical fiber polarization mode dispersion. Background Art

[0002] Optical fiber polarization mode dispersion (PMD) is a dispersion effect of optical fibers. In an optical fiber communication system, due to the random birefringence effect of the optical fiber, the transmission rates of optical signals in two eigen - orthogonal polarization states are different, resulting in a time - delay difference, which is generally represented by differential group delay. This time - delay difference will limit the transmission bandwidth and distance of the optical fiber communication system and affect the system performance.

[0003] The generation of optical fiber polarization mode dispersion is mainly related to the birefringence effect of the optical fiber, and the birefringence effect is caused by the imperfections of the optical fiber, including: geometric asymmetry of the optical fiber, such as the ellipticity of the optical fiber, non - uniformity of the core diameter, etc.; anisotropy of the material, such as the crystal structure or non - uniform doping of the optical fiber material; external factors: such as bending, twisting, temperature change of the optical fiber, etc.

[0004] The basic principle of the distributed measurement technology of optical fiber polarization mode dispersion is that the polarization state of the optical signal will change randomly due to the random birefringence of the optical fiber. Currently, the methods applied to the distributed measurement of optical fiber polarization mode dispersion mainly include the Jones matrix analysis method and the pulse - width scanning method.

[0005] The Jones matrix analysis method uses an optical fiber - glass - plate model to equivalent the optical fiber as a cascade of several Jones matrices, and realizes the distributed measurement of optical fiber polarization mode dispersion through the Jones matrix analysis of the polarization - light time - domain reflection signal. Its main advantage is that it can obtain the specific value of the distributed optical fiber polarization mode dispersion, and the measurement accuracy is relatively high; its main disadvantage is that the polarization - state detection device needs to be adjusted successively to three linearly polarized lights and a right - hand circularly polarized light to obtain the Jones vector of the received optical signal, and the measurement steps are cumbersome.

[0006] The pulse - width scanning method uses the relationship between the pulse width of the optical signal and the beat length of the optical fiber, and estimates the distributed optical fiber polarization mode dispersion by analyzing the fluctuation of the polarization - light time - domain reflection curve by changing the optical pulse width. Its main advantage is that the measurement steps are simple and the equipment system is easy to install; its main disadvantage is that it cannot obtain the specific value of the distributed optical fiber polarization mode dispersion, and can only obtain the estimated interval measured by pulse - width scanning, and the measurement accuracy is low.

[0007] In addition, a patent document with the publication number "CN117134822A" provides an optical fiber polarization mode dispersion test system and method, including: a broadband light source; a polarizer; an analyzer; a polarization beam splitter for receiving the composite light from the output end of an optical fiber coupler and separating the composite light into a first interference pattern and a second interference pattern in two mutually orthogonal polarization states; a photodetector for converting the first interference pattern into a first interference pattern electrical signal or converting the second interference pattern into a second interference pattern electrical signal; a control and processing unit for processing the first interference pattern electrical signal and / or the second interference pattern electrical signal, calculating independent autocorrelation functions and cross-correlation functions, and determining the polarization mode dispersion of the device under test based on the autocorrelation functions and cross-correlation functions; however, this prior art still requires changing the polarization state of the optical signal at the receiving end multiple times during the measurement process, resulting in low measurement efficiency.

[0008] Currently, there is no solution that can achieve simple and efficient measurement of the accurate value of optical fiber polarization mode dispersion without changing the polarization state of the optical signal at the receiving end multiple times. Summary of the Invention

[0009] To overcome the defects of low efficiency and accuracy in measuring optical fiber polarization mode dispersion in the above prior art, the present invention provides a method and system for distributed measurement of optical fiber polarization mode dispersion, which can achieve simple and efficient measurement of the accurate value of distributed polarization mode dispersion of optical fiber without changing the polarization state of the optical signal at the receiving end multiple times.

[0010] To solve the above technical problems, the technical solution of the present invention is as follows:

[0011] A method for distributed measurement of optical fiber polarization mode dispersion includes the following steps:

[0012] Inject pulsed polarized light into an optical fiber under test with unknown parameters for transmission. Due to the birefringence effect, the polarization state of the pulsed polarized light changes randomly at different positions in the optical fiber under test during transmission.

[0013] Detect the polarization state of the Rayleigh scattered light of the pulsed polarized light and construct a polarization optical time domain reflectometry curve.

[0014] Extract the effective features of the polarization optical time domain reflectometry curve and input them into a pre-trained machine learning model to obtain the predicted result of the polarization mode dispersion of the optical fiber under test, realizing distributed polarization mode dispersion measurement.

[0015] Preferably, after polarization state detection, the change in the polarization state of the pulsed polarized light is converted into a change in optical power, expressed as:

[0016]

[0017] Wherein, T(z) is the received optical power; R(z) is the Jones matrix from the emission end of the pulsed polarized light to the Rayleigh scattering point z in the equivalent glass plate model of the optical fiber; M is the reflection Jones matrix; is the Jones vector of the pulsed polarized light input to the optical fiber to be measured; is the Jones vector of the polarization state of the pulsed polarized light after polarization state detection;

[0018] Based on the change of the optical power, a polarization optical time domain reflectometry (P-OTDR) curve is constructed.

[0019] Preferably, a pre-trained machine learning model is obtained according to the following steps:

[0020] Obtain several optical fibers with known parameters, and calculate the differential group delay of each optical fiber;

[0021] Inject the pulsed polarized light into each optical fiber for transmission respectively, detect the polarization state of the Rayleigh scattered light of the pulsed polarized light respectively, and construct a P-OTDR curve respectively. Extract the effective features of each P-OTDR curve, and the effective features at least include the curve level crossing rate, slope and average power;

[0022] Take the effective features of the P-OTDR curves corresponding to the optical fibers with known parameters, and their calculated differential group delays as the training sample set;

[0023] Construct a machine learning model, and use the training sample set to train the machine learning model to obtain a pre-trained machine learning model.

[0024] Preferably, the polarization mode dispersion (PMD) of the optical fiber is represented by the differential group delay, and the differential group delay of the optical fiber with known parameters is calculated according to the following formula:

[0025]

[0026] Wherein, <Δτ> is the average value of the differential group delays of multiple optical fiber samples with the same length, λ is the wavelength of the pulsed polarized light, c is the speed of light in vacuum, L B is the beat length of the optical fiber, L F is the coupling length of the optical fiber, and L is the total length of the optical fiber.

[0027] Preferably, the machine learning model is specifically any one of a support vector machine, a random forest regression model and a deep neural network.

[0028] The present invention also provides an optical fiber polarization mode dispersion distributed measurement system, which applies the above-mentioned optical fiber polarization mode dispersion distributed measurement method, and includes:

[0029] A pulsed polarized light generation module for generating pulsed polarized light;

[0030] The optical fiber to be measured is used to transmit the pulsed polarized light;

[0031] The polarization state detection module is used to detect the polarization state of the Rayleigh scattered light of the pulsed polarized light and obtain the polarization optical time domain reflectance curve;

[0032] The circulator, the first end of the circulator is connected to the pulsed polarized light generation module, the second end of the circulator is connected to the optical fiber to be measured, and the third end of the circulator is connected to the polarization state detection module;

[0033] The data processing module is used to extract the effective features of the polarization optical time domain reflectance curve collected by the polarization state detection module and use the pre-trained machine learning model to predict the polarization mode dispersion of the optical fiber to be measured.

[0034] Preferably, the pulsed polarized light generation module includes, connected in sequence: a light source, a polarization controller, and a pulse modulator. The light generated by the light source forms polarized light after passing through the polarization controller, and the polarized light forms pulsed polarized light after passing through the pulse modulator.

[0035] Preferably, the pulsed polarized light generation module includes, connected in sequence: a pulsed laser and a polarization controller. The pulsed laser is used to generate pulsed light, and the pulsed light forms pulsed polarized light after passing through the polarization controller.

[0036] Preferably, the polarization state detection module includes, connected in sequence: a polarization state detection device, a photodetector, and a signal acquisition card. After the polarization state detection device detects the pulsed polarized light returned by Rayleigh scattering, it is converted into an electrical signal by the photodetector and transmitted to the signal acquisition card. The signal acquisition card stores the collected signal in the data processing module.

[0037] Preferably, the data processing module includes: a feature extraction sub-module and a machine learning prediction sub-module;

[0038] The feature extraction sub-module is used to extract the effective features of the polarization optical time domain reflectance curve. The effective features at least include the curve level crossing rate and the slope;

[0039] The machine learning prediction sub-module is used to predict the differential group delay of the optical fiber to be measured according to the extracted effective features, use the pre-trained machine learning model to obtain the polarization mode dispersion prediction result of the optical fiber to be measured, and realize distributed polarization mode dispersion measurement.

[0040] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0041] The present invention provides a method and system for distributed measurement of polarization mode dispersion in optical fibers. First, a pulsed polarized light is injected into an optical fiber under test with unknown parameters for transmission. Due to the birefringence effect, the polarization state of the pulsed polarized light randomly changes at different positions in the optical fiber under test during transmission. Then, the polarization state of the Rayleigh scattered light of the pulsed polarized light is detected, and a polarization optical time domain reflectometry (P-OTDR) curve is constructed. Finally, the effective features of the P-OTDR curve are extracted and input into a pre-trained machine learning model to obtain the prediction result of the polarization mode dispersion of the optical fiber under test, realizing distributed polarization mode dispersion measurement.

[0042] Compared with the prior art, the present invention has the following advantages:

[0043] 1) Compared with the Jones matrix analysis method, the fiber distributed polarization mode dispersion measurement method of the present invention does not require multiple changes in the polarization state of the received optical signal, and the measurement steps are simple and efficient, and the system is easy to install.

[0044] 2) Compared with the pulse width scanning method, the fiber distributed polarization mode dispersion measurement method of the present invention is based on a machine learning model, can accurately measure the fiber distributed polarization mode dispersion value, has high measurement accuracy, and does not increase the complexity of the measurement system and method, and has high measurement efficiency. Description of the Drawings

[0045] Figure 1 It is a flowchart of a method for distributed measurement of polarization mode dispersion in an optical fiber provided in Embodiment 1.

[0046] Figure 2 It is a schematic diagram of the OTDR Rayleigh scattering principle provided in Embodiment 1.

[0047] Figure 3 It is a schematic diagram of a measurement structure for obtaining the total differential group delay value of an optical fiber based on the spectral interference method provided in Embodiment 1.

[0048] Figure 4 It is a spectrogram of a 25.115 km optical fiber measured based on the spectral interference method provided in Embodiment 1.

[0049] Figure 5 It is a schematic diagram of a measurement structure for obtaining the distributed beat length of an optical fiber based on a polarization optical time domain reflectometer provided in Embodiment 1.

[0050] Figure 6 It is a result diagram of the distributed beat length of a 25.115 km optical fiber measured based on a polarization optical time domain reflectometer provided in Embodiment 1.

[0051] Figure 7 It is a result diagram of calculating the distributed polarization mode dispersion of a 25.115 km optical fiber based on the fiber coupling length and the distributed beat length provided in Embodiment 1.

[0052] Figure 8 It is the overall working flow chart of a fiber optic polarization mode dispersion distributed measurement system provided in Embodiment 2.

[0053] Figure 9 It is the hardware structure diagram of a fiber optic polarization mode dispersion distributed measurement system provided in Embodiment 2.

[0054] Figure 10 It is another hardware structure diagram of a fiber optic polarization mode dispersion distributed measurement system provided in Embodiment 2. Specific implementation manners

[0055] The attached drawings are only for illustrative purposes and cannot be construed as a limitation to this patent;

[0056] To better illustrate this embodiment, some components in the attached drawings will be omitted, enlarged or reduced, which do not represent the dimensions of the actual product;

[0057] For those skilled in the art, it is understandable that some well-known structures and their descriptions in the attached drawings may be omitted.

[0058] The technical solution of the present invention will be further described below in conjunction with the attached drawings and embodiments.

[0059] Embodiment 1

[0060] As Figure 1 shown, this embodiment provides a fiber optic polarization mode dispersion distributed measurement method, including the following steps:

[0061] S1: Inject pulsed polarized light into the fiber under test with unknown parameters for transmission. Due to the birefringence effect, the polarization state of the pulsed polarized light changes randomly at different positions in the fiber under test during transmission;

[0062] S2: Detect the polarization state of the Rayleigh scattered light of the pulsed polarized light and construct a polarization optical time domain reflectometry curve;

[0063] S3: Extract the effective features of the polarization optical time domain reflectometry curve and input them into a pre-trained machine learning model to obtain the polarization mode dispersion prediction result of the fiber under test, realizing distributed polarization mode dispersion measurement;

[0064] After polarization state detection, the change in the polarization state of the pulsed polarized light is converted into a change in optical power, expressed as:

[0065]

[0066] where, T(z) is the received optical power; R(z) is the Jones matrix from the pulsed polarized light emission end to the Rayleigh scattering point z in the fiber equivalent retarder model; M is the reflection Jones matrix; is the Jones vector of the pulsed polarized light input to the optical fiber to be measured; is the Jones vector of the polarization state of the pulsed polarized light after polarization state detection;

[0067] Construct a polarization optical time domain reflectometry curve based on the change of the optical power;

[0068] Obtain a pre-trained machine learning model according to the following steps:

[0069] Obtain several optical fibers with known parameters, and calculate the differential group delay of each optical fiber;

[0070] Use the differential group delay to represent the polarization mode dispersion of the optical fiber. The differential group delay of the optical fiber with known parameters is calculated according to the following formula:

[0071]

[0072] where <Δτ> is the average value of the differential group delays of multiple optical fiber samples of the same length, λ is the wavelength of the pulsed polarized light, c is the speed of light in vacuum, L B is the beat length of the optical fiber, L F is the coupling length of the optical fiber, and L is the total length of the optical fiber;

[0073] Inject the pulsed polarized light into each optical fiber for transmission, respectively detect the polarization state of the Rayleigh scattered light of the pulsed polarized light, and respectively construct a polarization optical time domain reflectometry curve, and extract the effective features of each polarization optical time domain reflectometry curve. The effective features at least include the curve level crossing rate and the slope;

[0074] Take the effective features of the polarization optical time domain reflectometry curves corresponding to each optical fiber with known parameters, and the calculated differential group delay as a training sample set together;

[0075] Construct a machine learning model, and use the training sample set to train the machine learning model to obtain a pre-trained machine learning model;

[0076] In this embodiment, the machine learning model is specifically any one of a support vector machine, a random forest regression model, and a deep neural network.

[0077] In the specific implementation process, first inject the pulsed polarized light into the optical fiber to be measured with unknown parameters for transmission; due to the birefringence effect, the polarization state of the pulsed polarized light changes randomly at different positions of the optical fiber to be measured during transmission;

[0078] Illustrate the basic concept and measurement principle of fiber polarization mode dispersion (PMD) based on the relationship between PMD and fiber beat length and coupling length; due to factors such as inhomogeneous fiber material, asymmetric structure, and strain, the random birefringence effect causes the transmission rates of optical signals in two intrinsically orthogonal polarization states to be different, resulting in a time delay difference, which is generally expressed by differential group delay as follows:

[0079]

[0080] where <Δτ> is the mean value of differential group delay of multiple fiber samples of the same length. Since the differential group delay of different fiber samples of the same type and length usually varies due to fiber random birefringence, the mean value is used here to represent it; λ is the wavelength of the optical signal, c is the speed of light in vacuum, L B is the fiber beat length, L F is the fiber coupling length, and L is the total length of the fiber;

[0081] Optical time domain reflectance (OTDR) signals mainly follow the Rayleigh scattering principle; as Figure 2 shown, the main cause of Rayleigh scattering is the inhomogeneity of the fiber itself, which is an elastic scattering; the frequency of the scattered light of Rayleigh scattering is equal to the frequency of the incident signal light, and Rayleigh scattering itself does not cause a sudden change in the polarization state of light; positioning based on the relationship between the propagation time and distance of the scattered light is the most basic positioning principle of distributed fiber optic sensing based on OTDR technology;

[0082] If an optical pulse propagates in the fiber at a speed of v, assuming the time when the optical pulse is sent is 0, after a time of t, the receiver just receives the Rayleigh scattering signal from point z (the distance of point z from the optical pulse launch end is z / 2), z can be expressed in terms of v and t as follows:

[0083]

[0084] where c is the speed of light in vacuum and n is the refractive index of the fiber core;

[0085] There is loss when light propagates in the fiber. The total loss of the Rayleigh scattering signal from point z (compared to when the optical pulse is just sent out) is:

[0086] P(z) = P0e -αz (3)

[0087] where P0 is the initial optical power when the optical pulse enters the fiber, and α is the loss coefficient of the fiber; for a pulsed light with a pulse width of T, the backscattered Rayleigh scattering optical power near point z is PB(z) as shown in the following equation:

[0088]

[0089] where α Ris the Rayleigh scattering coefficient of the optical fiber, β R is the backscattering coefficient;

[0090] Detect the polarization state of the Rayleigh scattered light of the detection pulse polarized light, and construct a polarization optical time domain reflectance curve;

[0091] In this embodiment, the change in the polarization state of the pulsed polarized light is converted into a change in optical power at the receiving end by a polarization detection device, which is expressed as:

[0092]

[0093] where R(z) is the Jones matrix from the optical pulse signal transmitting end to the Rayleigh scattering point z in the fiber equivalent slab model, M is the reflection Jones matrix, is the Jones vector of the pulsed polarized light input to the optical fiber to be measured, is the Jones vector of the polarization state of the optical signal passing through the polarization state detection device; construct a polarization optical time domain reflectance curve based on the change in optical power;

[0094] After that, extract the effective features of the polarization optical time domain reflectance curve and input them into a pre-trained machine learning model to obtain the polarization mode dispersion prediction result of the optical fiber to be measured, realizing distributed polarization mode dispersion measurement;

[0095] In this embodiment, the pre-trained machine learning model is obtained according to the following steps:

[0096] Obtain several optical fibers with known parameters, and calculate the differential group delay of each optical fiber;

[0097] For the specific calculation process, as Figure 3 shown, first, the differential group delay value of the entire optical fiber can be measured by the spectral interference method, and the measurement result is as Figure 4 shown, and the specific calculation formula is:

[0098]

[0099] where λ1 is the wavelength corresponding to the first interference peak in the spectrum except for the intrinsic peak of the erbium-doped fiber amplifier, λ N is the wavelength corresponding to the Nth interference peak in the spectrum except for the intrinsic peak of the erbium-doped fiber amplifier, c is the speed of light in vacuum, and N is the number of interference periods between two interference peaks;

[0100] The distributed beat length of the optical fiber can be measured by using a polarization optical time domain reflectometer, and the measurement structure is as Figure 5 shown, and the specific calculation formula is:

[0101]

[0102] Among them, ν is a certain light intensity value on the normalized polarization optical time domain reflectance curve, with a value range of 0 to 1, n(ν) is the number of times the curve crosses the ν value per unit length of the polarization optical time domain reflectance curve, and the measurement results are as Figure 6 shown;

[0103] The average beat length of the optical fiber can be calculated from the distributed beat length measurement results. Substituting the total differential group delay and the average beat length of the optical fiber into Equation (1) can obtain the coupling length of the optical fiber. Then, using the coupling length and the distributed beat length, the distributed polarization mode dispersion value of the optical fiber can be calculated from Equation (1), and the calculation results are as Figure 7 shown;

[0104] That is, Equation (1) is used to calculate the distributed differential group delay as the standard value for training, and Equation (6) is used to calculate the total differential group delay of the optical fiber, so as to obtain parameters such as the beat length and the coupling length. Furthermore, Equation (1) is used to obtain the standard values of the differential group delay of each optical fiber for training the machine learning model;

[0105] In this method, the results calculated by the above formulas are used to train the machine learning model. Finally, for the polarization mode dispersion value of the optical fiber to be measured, the result predicted by the trained machine learning model shall prevail;

[0106] Inject the pulsed polarized light into each optical fiber for transmission respectively, detect the polarization states of the Rayleigh scattered light of the pulsed polarized light respectively, and construct the polarization optical time domain reflectance curves respectively. Extract the effective features of each polarization optical time domain reflectance curve. The effective features at least include the curve level crossing rate, the slope, and the average power;

[0107] The effective features of the polarization optical time domain reflectance curves corresponding to the optical fibers with known parameters respectively, and the differential group delay calculated therefrom are jointly used as the training sample set;

[0108] Construct a machine learning model, and use the training sample set to train the machine learning model to obtain a pre-trained machine learning model;

[0109] In this embodiment, the machine learning model can be any one of a support vector machine, a random forest regression model, and a deep neural network, or other conventional models. This embodiment does not limit this here;

[0110] This method can achieve simple and efficient measurement of the accurate value of the distributed polarization mode dispersion of the optical fiber without the need to change the polarization state of the optical signal at the receiving end multiple times.

[0111] Embodiment 2

[0112] This embodiment provides a distributed measurement system for the polarization mode dispersion of an optical fiber, which applies the distributed measurement method for the polarization mode dispersion of an optical fiber described in Embodiment 1, and includes:

[0113] A pulsed polarized light generation module for generating pulsed polarized light;

[0114] A fiber optic to be measured for transmitting the pulsed polarized light;

[0115] A polarization state detection module for detecting the polarization state of the Rayleigh scattered light of the pulsed polarized light and obtaining a polarization optical time domain reflectometry curve;

[0116] An optical circulator, the first end of the optical circulator is connected to the pulsed polarized light generation module, the second end of the optical circulator is connected to the fiber optic to be measured, and the third end of the optical circulator is connected to the polarization state detection module;

[0117] A data processing module for extracting effective features of the polarization optical time domain reflectometry curve collected by the polarization state detection module and predicting the polarization mode dispersion of the fiber optic to be measured using a pre-trained machine learning model;

[0118] The pulsed polarized light generation module includes, connected in sequence: a light source, a polarization controller, and a pulse modulator. The light generated by the light source forms polarized light after passing through the polarization controller, and the polarized light forms pulsed polarized light after passing through the pulse modulator; or, the pulsed polarized light generation module includes, connected in sequence: a pulsed laser and a polarization controller. The pulsed laser is used to generate pulsed light, and the pulsed light forms pulsed polarized light after passing through the polarization controller;

[0119] The polarization state detection module includes, connected in sequence: a polarization state detection device, a photodetector, and a signal acquisition card. After the polarization state detection device detects the pulsed polarized light returned by Rayleigh scattering, it is converted into an electrical signal by the photodetector and transmitted to the signal acquisition card. The signal acquisition card stores the collected signal into the data processing module;

[0120] The data processing module includes: a feature extraction sub-module and a machine learning prediction sub-module;

[0121] The feature extraction sub-module is used to extract effective features of the polarization optical time domain reflectometry curve. The effective features at least include the curve level crossing rate and slope;

[0122] The machine learning prediction sub-module is used to predict the differential group delay of the fiber optic to be measured according to the extracted effective features, obtain the polarization mode dispersion prediction result of the fiber optic to be measured, and realize distributed polarization mode dispersion measurement.

[0123] In the specific implementation process, as Figure 8 shown in the overall process, the overall working process of this system is as follows:

[0124] First, the light source emits detection light, as Figure 9As shown in the hardware structure diagram, if the light source is a common light source, the continuous light emitted first passes through a polarization controller to become continuous polarized light and then passes through a pulse modulator to output pulsed polarized light; as Figure 10 As shown in the hardware structure diagram, if the light source is a pulsed laser, the pulsed light emitted becomes pulsed polarized light after passing through a polarization controller;

[0125] The pulsed polarized light enters the fiber under test through the first and second ports of the optical circulator for transmission. Due to the random birefringence effect caused by factors such as non-uniform fiber material, asymmetric structure, and strain, the polarization state of the pulsed polarized light changes randomly at different positions in the fiber under test during transmission;

[0126] Part of the detected light enters the polarization state detection device through the second and third ports of the optical circulator due to Rayleigh scattering everywhere in the fiber. The polarization state of the optical signal remains unchanged before and after Rayleigh scattering;

[0127] After that, the optical signal output by the polarization state detection device is collected by a photodetector and transmitted to a signal acquisition card, and the signal acquisition card stores the collected signal in a data processing module;

[0128] The data processing module includes two parts: a feature extraction sub-module and a machine learning prediction sub-module; the feature extraction sub-module is used to extract effective features of the polarization optical time domain reflectometry curve (such as curve level crossing rate, slope, and average power, etc.); the machine learning prediction sub-module is used to predict the differential group delay of the fiber under test based on the extracted effective features, using a pre-trained machine learning model, and obtain the polarization mode dispersion prediction result of the fiber under test to achieve distributed polarization mode dispersion measurement;

[0129] This system can achieve simple and efficient measurement of the accurate value of fiber distributed polarization mode dispersion without the need to change the polarization state of the received optical signal multiple times.

[0130] The same or similar reference numerals correspond to the same or similar components;

[0131] The terms describing the positional relationship in the drawings are only for illustrative purposes and should not be construed as a limitation of this patent;

[0132] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention and are not limitations on the embodiments of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the embodiments here. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims

1. A distributed measurement method for polarization mode dispersion of optical fibers, characterized in that It includes the following steps: Inject pulsed polarized light into the optical fiber under test with unknown parameters for transmission. Due to the birefringence effect, the polarization state of the pulsed polarized light changes randomly at different positions in the optical fiber under test during transmission; Detect the polarization state of the Rayleigh scattered light of the pulsed polarized light and construct a polarization optical time domain reflectometry (P-OTDR) curve; Extract the effective features of the P-OTDR curve and input them into a pre-trained machine learning model to obtain the polarization mode dispersion (PMD) prediction result of the optical fiber under test, realizing distributed PMD measurement.

2. The distributed measurement method of fiber optic polarization mode dispersion according to claim 1, wherein After polarization state detection, convert the change in the polarization state of the pulsed polarized light into a change in optical power, which is expressed as: Among them, T(z) is the received optical power; R(z) is the Jones matrix from the emission end of the pulsed polarized light to the Rayleigh scattering point z in the equivalent glass plate model of the optical fiber; M is the reflection Jones matrix; is the Jones vector of the pulsed polarized light input to the optical fiber under test; is the Jones vector of the polarization state of the pulsed polarized light after polarization state detection; Construct a P-OTDR curve based on the change in the optical power.

3. The distributed measurement method for fiber optic polarization mode dispersion according to claim 1, wherein Obtain a pre-trained machine learning model according to the following steps: Obtain several optical fibers with known parameters and calculate the differential group delay of each optical fiber; Inject pulsed polarized light into each optical fiber for transmission respectively, detect the polarization state of the Rayleigh scattered light of the pulsed polarized light respectively, and construct a P-OTDR curve respectively. Extract the effective features of each P-OTDR curve. The effective features at least include the curve level crossing rate, slope, and average power; Use the effective features of the P-OTDR curves corresponding to the optical fibers with known parameters and their calculated differential group delays as a training sample set together; Construct a machine learning model and use the training sample set to train the machine learning model to obtain a pre-trained machine learning model.

4. A distributed measurement method for fiber optic polarization mode dispersion according to claim 3, characterized in that, Use the differential group delay to represent the PMD of the optical fiber. The differential group delay of the optical fiber with known parameters is calculated according to the following formula: where <Δτ> is the mean differential group delay of multiple optical fiber samples of the same length, λ is the wavelength of the pulsed polarized light, c is the speed of light in vacuum, L B is the beat length of the optical fiber, L F is the coupling length of the optical fiber, and L is the total length of the optical fiber.

5. A distributed measurement method for fiber optic polarization mode dispersion according to any one of claims 1 to 4, characterized in that The machine learning model is specifically any one of a support vector machine, a random forest regression model, and a deep neural network.

6. A distributed measurement system for fiber optic polarization mode dispersion, which applies the distributed measurement method for fiber optic polarization mode dispersion described in any one of claims 1 to 5, characterized in that, It includes: A pulsed polarized light generation module for generating pulsed polarized light; An optical fiber under test for transmitting the pulsed polarized light; A polarization state detection module for detecting the polarization state of the Rayleigh scattered light of the pulsed polarized light and obtaining a P-OTDR curve; A circulator. The first end of the circulator is connected to the pulsed polarized light generation module, the second end of the circulator is connected to the optical fiber under test, and the third end of the circulator is connected to the polarization state detection module; A data processing module for extracting the effective features of the P-OTDR curve collected by the polarization state detection module and predicting the PMD of the optical fiber under test using a pre-trained machine learning model.

7. The distributed measurement system for fiber optic polarization mode dispersion according to claim 6, characterized in that, The pulsed polarized light generation module includes, connected in sequence: a light source, a polarization controller, and a pulse modulator. The light generated by the light source forms polarized light after passing through the polarization controller, and the polarized light forms pulsed polarized light after passing through the pulse modulator.

8. A distributed measurement system for fiber optic polarization mode dispersion according to claim 6, characterized in that The pulsed polarized light generation module includes, connected in sequence: a pulsed laser and a polarization controller. The pulsed laser is used to generate pulsed light, and the pulsed light forms pulsed polarized light after passing through the polarization controller.

9. A distributed measurement system for fiber optic polarization mode dispersion according to any one of claims 6 to 8, characterized in that, The polarization state detection module includes, connected in sequence: a polarization state detection device, a photodetector, and a signal acquisition card. After the polarization state detection device detects the pulsed polarized light returned by Rayleigh scattering, it is converted into an electrical signal by the photodetector and transmitted to the signal acquisition card, and the signal acquisition card stores the acquired signal into the data processing module.

10. A distributed measurement system for fiber optic polarization mode dispersion according to claim 9, characterized in that, The data processing module includes: a feature extraction sub-module and a machine learning prediction sub-module; The feature extraction sub-module is used to extract the effective features of the polarization light time-domain reflection curve, and the effective features at least include the curve level crossing rate and the slope; The machine learning prediction sub-module is used to predict the differential group delay of the optical fiber to be measured according to the extracted effective features, use a pre-trained machine learning model to obtain the polarization mode dispersion prediction result of the optical fiber to be measured, and realize distributed polarization mode dispersion measurement.

Images