CN-121994290-A - BOTDR or BOTDA temperature or strain error estimation method for high spatial resolution scene

CN121994290ACN 121994290 ACN121994290 ACN 121994290ACN-121994290-A

Abstract

The invention discloses a temperature or strain error estimation method of a Brillouin optical time domain reflectometer (Brillouin optical time domain reflectometer, BOTDR) or Brillouin optical time domain analysis (Brillouin optical time domain analysis, BOTDA) aiming at a high-spatial resolution scene. Aiming at a scene with high spatial resolution, the invention carries out linearization processing on residual errors through first-order Taylor expansion, and derives and obtains a Gauss-Newton normal equation of a Gaussian spectrum fitting method. Further, by calculating covariance among parameters to be fitted, a brillouin frequency shift (Brillouin frequency shift BFS) error estimation formula under high spatial resolution is obtained through first derivation. Based on the formula, BFS error can be estimated when the Brillouin line width, the Brillouin gain peak value, the frequency sweep interval, the frequency sweep point number, the noise standard deviation and the noise level are known, and further, temperature or strain error of BOTDR or BOTDA can be estimated conveniently. The method fills the blank of temperature or strain error estimation of BOTDR or BOTDA under high spatial resolution, and provides reference value for improving the accuracy of BOTDR or BOTDA.

Inventors

ZHAO LIJUAN
CHEN YONGHUI
XU ZHINIU

Assignees

华北电力大学（保定）

Dates

Publication Date: 20260508
Application Date: 20260203

Claims (4)

1. A brillouin optical time domain reflectometer (Brillouin optical time domain reflectometer, BOTDR) or brillouin optical time domain analyzer (Brillouin optical time domain analysis, BOTDA) temperature or strain error estimation method for high spatial resolution scenarios, characterized in that the method estimates the brillouin frequency shift (Brillouin frequency shift, BFS) error, and hence the BOTDR or BOTDA temperature or strain error, from the brillouin linewidth, the brillouin gain peak, the sweep interval, the number of sweep points, the noise standard deviation and the noise level.
2. A method for estimating BOTDR or BOTDA temperature or strain error for high spatial resolution scenarios in accordance with claim 1, wherein the BFS error calculation formula proposed by the present invention is as follows: Wherein sigma B is a Brillouin frequency shift error, sigma M is a standard deviation of gain of a frequency sweep point closest to the Brillouin frequency shift, and g F and Deltav BF are respectively the average value of a Brillouin gain peak value and a Brillouin line width obtained by fitting noisy Brillouin spectra under a plurality of same parameters. Q 1 and Q 2 are coefficients to be calculated, calculated by: wherein delta is the frequency sweep interval, alpha c is the noise figure factor, and N is the frequency sweep point number.
3. A BOTDR or BOTDA temperature or strain error estimation method for high spatial resolution scenarios in accordance with claim 2, wherein said step of obtaining said noise figure factor α c comprises the steps of: 1) Acquiring Brillouin spectra in the sensing optical fiber by using BOTDR or BOTDA, and repeating the acquisition for ten times; 2) Fitting the actually measured brillouin spectrum by using Gaussian spectrum fitting method to obtain brillouin spectrum coefficient, namely ; 3) Reconstructing a noise-free brillouin spectrum by using the brillouin spectral coefficient to obtain Subtracting the measured Brillouin spectrum from the measured Brillouin spectrum to obtain noise; 4) Taking standard deviation of 10 groups of noise under each space point to obtain the standard deviation of the noise under each frequency point, and carrying out average treatment on the standard deviation of the noise to obtain sigma i ; 5) Averaging the brillouin spectrum coefficients under all repetition, and reconstructing a brillouin spectrum by using the brillouin spectrum coefficients to obtain f i (a F ); 6) According to The noise standard deviation is fitted, where α c and β c are the noise figure factor and the scale factor, respectively.
4. A method of BOTDR or BOTDA temperature or strain error estimation for high spatial resolution scenarios in accordance with claim 1, wherein the formula for calculating the temperature or strain error from BFS error is as follows: Wherein sigma T and sigma ε are errors of temperature and strain respectively, C T and C ε are temperature sensitivity coefficients and strain sensitivity coefficients of Brillouin frequency shift respectively, and delta T and delta epsilon are variation of temperature and strain respectively.

Description

BOTDR or BOTDA temperature or strain error estimation method for high spatial resolution scene Technical Field The invention relates to a BOTDR or BOTDA temperature or strain error estimation method for a high-spatial resolution scene, and belongs to the technical field of measurement. Background The distributed optical fiber sensor has the characteristics of long-distance continuous monitoring, distributed measurement, strong electromagnetic interference resistance and adaptation to severe environments, and is widely applied to the fields of energy pipe networks, submarine cables, traffic infrastructures, safety monitoring and the like. Compared with the traditional point type electronic sensor, the multi-parameter and full-line distributed measurement can be realized by utilizing a single optical fiber, and the monitoring efficiency and the system reliability are remarkably improved. The brillouin optical time domain reflectometer (Brillouin optical time domain reflectometer, BOTDR)/brillouin optical time domain analyzer (Brillouin optical time domain analysis, BOTDA) can monitor the temperature and strain of the optical fiber. The difference between the center frequency of the brillouin spectrum (Brillouin gain spectrum, BGS) and the frequency of the incident light is referred to as the brillouin shift (Brillouin frequency shift, BFS), and the amount of change in the brillouin shift is linearly related to the amount of change in temperature or strain. Thus, the implementation of brillouin sensing relies on accurate extraction of BFS in BGS. The existing common BFS extraction method comprises a spectrum fitting method, a neural network method, a quadratic polynomial fitting method and the like. The neural network method and the quadratic polynomial fitting method have the advantages of high calculation speed, but the two methods have insufficient anti-interference capability, have higher requirements on the spectrum to be fitted, and have lower accuracy than the spectrum fitting method. For example, when the BGS spectrum of the test set is greatly different from that of the training set, the accuracy of the neural network method can be affected, and when BFS deviates from the center of the sweep frequency range, the accuracy of the quadratic polynomial fitting method can be greatly reduced. In contrast, the spectrum fitting method has extremely high accuracy and robustness. As the length of the fiber increases and the requirement for high spatial resolution (reduced pulse width of the incident light) the signal to noise ratio of the brillouin spectrum decreases, and it is not sufficient to measure only the temperature and strain, and the error estimation problem becomes important. The prior related patent is as follows, and the prior patent document with publication number of CN113155170A, namely a Brillouin frequency shift error estimation method, provides a quick and accurate Brillouin frequency shift error estimation method. The method comprises the following steps of injecting pulse light into an optical fiber, detecting a brillouin spectrum of scattered light at an incidence end of the optical fiber, fitting the actually measured brillouin spectrum by adopting a fitting algorithm based on a pseudo Voigt model, and calculating to obtain a line width and a signal to noise ratio in the line width. The Brillouin frequency shift error can be conveniently, rapidly and accurately calculated according to the proposed formula. Meanwhile, the influence rule of each factor on the error can be qualitatively analyzed according to the estimation formula, and a reference is provided for selecting proper parameters during actual measurement to further reduce the error of the Brillouin frequency shift. The invention discloses an error estimation method for extracting Brillouin frequency shift of a neural network, which is disclosed in the prior patent with the publication number of CN119492406A, and aims at the neural network with an activation function being a linear function, and the invention obtains the relation between the variance of each node of an input layer of the neural network and the standard deviation of output of the neural network. The standard deviation of the Brillouin frequency shift calculated by the neural network can be obtained through the extracted neural network weight matrix and the variance of the Brillouin gain under each frequency of the Brillouin spectrum. The method fills the blank of neural network error estimation with the activation function being a linear function, and can further improve the accuracy of the neural network in extracting the Brillouin frequency shift. The above method describes two brillouin frequency shift error estimation methods, but the formula in the patent 'brillouin frequency shift error estimation method' is obtained by fitting a large number of simulated brillouin spectra, and if a noise model or a spectrum model changes, the error of the formula may increas