A method for improving the spatial resolution of distributed optical fiber sensing systems

Abstract

The present invention relates to the field of optical fiber sensing and detection technology, and in particular to a method for improving the spatial resolution of a distributed optical fiber sensing system. First, a reference signal and a test signal containing strain are collected, a two-dimensional image signal is generated through Fourier transform and cross-correlation calculation, and a time series analysis model is used for denoising; a global-local hybrid search strategy is used to dynamically optimize hyperparameters; and finally, a high-resolution strain distribution result is output to eliminate outliers and improve measurement stability. The present invention solves the defect that traditional denoising algorithms are not sufficiently adaptable to dynamic noise, and can be extended to distributed optical fiber sensing scenarios such as temperature and vibration, providing reliable support for high-precision monitoring.

Description

Technical Field

[0001] The present invention relates to the field of optical fiber sensing and detection technology, and in particular to a method for improving the spatial resolution of a distributed optical fiber sensing system. Background Art

[0002] In distributed fiber optic sensing technology, optical fiber is used not only as a sensing medium, but also as a transmission medium for measuring signals. By utilizing the transmission characteristics of light waves in optical fibers, such as Raman scattering, Rayleigh scattering, and Brillouin scattering, real-time monitoring of the external environment along the length of the optical fiber can be achieved. This technology has significant advantages such as strong anti-electromagnetic interference ability, simple structure, high spatial resolution, and long sensing distance. Based on these characteristics, distributed fiber optic sensing technology has been gradually applied to many fields, including bridge safety monitoring, civil engineering inspection, underground tunnel fire alarm, and geological survey, and has played an important social benefit. As a representative of this technology, optical frequency domain reflectometry (OFDR) has the characteristics of lightness, small size, high sensitivity, strong anti-electromagnetic interference, and excellent spatial resolution. It can continuously measure the changes of external physical quantities such as strain, vibration, and temperature along the optical fiber.

[0003] The working principle of OFDR is: the linear frequency sweep light emitted by the tunable laser light source is divided into two beams through a coupler, one of which enters the optical fiber to be tested, and after returning through the back Rayleigh scattered light of the optical fiber to be tested, it interferes with the other reference light. By collecting the beat signal and performing fast Fourier transform processing, the distance domain information along the optical fiber can be obtained. In the measurement process of OFDR, it is necessary to first collect a reference signal without external influences, and then collect a test signal affected by the outside world, and calculate the external changes through cross-correlation operations. A prominent feature of the OFDR system is its high spatial resolution, which can reach the millimeter level, which makes it widely used in high-precision monitoring fields such as aerospace. However, when the spatial resolution is improved, the cross-correlation between the reference signal and the test signal will be significantly reduced, resulting in multiple peaks or false peaks in the cross-correlation results, and accurate measurement results cannot be obtained. Therefore, how to effectively improve the spatial resolution of the OFDR system has become an important research issue.

[0004] To this end, the present application designs a method for improving the spatial resolution of a distributed optical fiber sensing system to solve the above problems. Summary of the invention

[0005] In order to remedy the problems in the prior art, the present invention provides a method for improving the spatial resolution of a distributed optical fiber sensing system, by using a timing analysis model to denoise and optimize OFDR measurement data, and automatically adjusting timing analysis hyperparameters in combination with an optimal solution processing algorithm, thereby solving the multi-peak and false peak problems of traditional OFDR systems in high spatial resolution measurements. This method can effectively improve the OFDR system's ability to detect changes in physical quantities in high-resolution measurements, reduce errors caused by reduced correlation between reference signals and test signals, ensure that external changes can still be accurately measured at high spatial resolution, and thus improve the application effect of OFDR technology in the field of high-precision monitoring.

[0006] A method for improving the spatial resolution of a distributed optical fiber sensing system, characterized by comprising the following steps: S1, collect reference signal and test signal containing strain; S2, perform FFT mapping of the signal to the distance domain and divide it into windows, and calculate the cross-correlation result; S3, reconstructs the 2D image and denoises it using a time series analysis model; S4, uses a global-local hybrid search strategy to optimize model hyperparameters; S5, output high-resolution strain measurement results.

[0007] Furthermore, in order to better implement the present invention, the S1 is specifically to collect signals twice, one is a signal without strain information, which is a reference signal; the other is a signal containing strain information, which is a test signal.

[0008] Further, in order to better implement the present invention, S2 is specifically, S21, mapping the reference signal and the test signal to the distance domain through fast Fourier transform, and taking a window size A to divide the signal into N equal parts, where the window size A determines the spatial resolution of the system; S22, performing a fast inverse Fourier transform on the local distance domain information of the first reference signal and the measurement signal; S23, performing cross-correlation calculation on the reference signal and the measurement signal after the fast inverse Fourier transform to obtain a one-dimensional cross-correlation result.

[0009] Further, in order to better implement the present invention, the S3 is specifically, S31, repeating S22-S23 to obtain the cross-correlation result of each corresponding position of the optical fiber, reconstructing all the obtained one-dimensional cross-correlation results into a two-dimensional image signal at the optical fiber distance, and performing denoising on the image based on the two-dimensional image by using a time series analysis-optimal solution analysis processing method; S32, optimizing the hyperparameters of the time series analysis model by an optimal solution analysis and processing method, and obtaining the optimal hyperparameters, wherein the hyperparameters are the number of model units and the Dropout ratio. These hyperparameters are used as inputs to reconstruct and train the time series analysis model, so as to obtain measurement results at a high spatial resolution and improve the accuracy of the measurement; S33, in the time series analysis model of the processing stage, the time series analysis model will predict the new data according to the rules learned in the training process, remove the noise in the original signal, add the Dropout layer to prevent overfitting, and regularize the model through the noise in the training data; use the mean square error MSE as the loss function, and continuously adjust the model parameters through the training process; in, , n is the total number of samples, is the true value of the i-th sample, is the predicted value of the i-th sample.

[0010] Furthermore, in order to better realize the present invention, the S4 is specifically, in the S32 stage, the hyperparameters in the timing analysis model are tuned using the optimal solution analysis and processing method, and a strategy combining global and local searches is adopted to automatically adjust the hyperparameters in the timing analysis model. Through multiple iterations, the performance of each parameter combination is evaluated, and finally the optimal hyperparameter configuration is found, thereby minimizing the loss of the model on the training set.

[0011] Furthermore, in order to better implement the present invention, the S5 is specifically to determine the appropriate spectrum shift between the reference signal and the measurement signal through cross-correlation analysis, and analyze the strain information along the sensing optical fiber.

[0012] The beneficial effects of the present invention are: When processing time series data, the present invention can effectively capture the long-term dependencies in the data, and can better model and predict the fluctuations and periodic changes that may occur in the OFDR system, thereby improving the resolution.

[0013] The present invention can more accurately identify tiny signal differences, especially in long-distance measurement, can improve spatial resolution, and reduce errors caused by noise and signal attenuation.

[0014] The present invention can further improve the processing accuracy of the optical frequency domain reflection system. The optimized model can work stably in complex environments, which significantly improves the spatial resolution of the system.

[0015] The technical solution of the present invention is not only applicable to traditional OFDR systems, but can also be extended to various distributed sensor systems, such as intelligent optical fiber sensing, structural health monitoring, temperature sensing and other fields. In these applications, the high resolution and high precision characteristics of the present invention can provide more accurate data support for measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a distributed optical fiber sensing system device based on a timing analysis-optimal solution analysis and processing method used in the present invention; Figure 2 is a flow chart of the present invention; Figure 3 This is a graph of the sensing fiber spectrum shift results obtained by the existing traditional processing method; Figure 4 This is a diagram of the sensing optical fiber spectrum deviation results obtained by the processing method of the present invention. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Generally, the components of the embodiments of the present invention described and shown in the accompanying drawings here can be arranged and designed in various different configurations.

[0018] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0019] This embodiment improves the distributed spatial resolution of the OFDR system by combining the timing analysis-optimal solution analysis and processing method. By analyzing the relationship between the reference signal and the measurement signal, optimizing the timing analysis model parameters, effectively removing noise, and enhancing measurement accuracy, the optimal solution analysis and processing method solves the problem of reduced correlation between the reference signal and the test signal in high-resolution measurement by automatically adjusting the hyperparameters of the timing analysis, ensuring accurate measurement of external change information.

[0020] like Figure 2 As shown, a measurement method for improving the distributed spatial resolution of an OFDR system based on a time series analysis-optimal solution analysis and processing method includes the following steps: 1. Collect signals twice, one without strain information, which is the reference signal; the other with strain information, which is the test signal; 2. Map the reference signal and the test signal to the distance domain through fast Fourier transform, and divide the signal into N equal parts using a window size A, where the window size A determines the spatial resolution of the system; 3. Perform fast inverse Fourier transform on the local distance domain information of the first reference signal and the measurement signal; 4. Obtain a one-dimensional cross-correlation result by cross-correlation calculation of the reference signal and the measurement signal after the fast inverse Fourier transform; 5. Repeat 3-4 to obtain the cross-correlation result of each corresponding position of the optical fiber, reconstruct all the one-dimensional cross-correlation results into a two-dimensional image signal at the optical fiber distance, and denoise the image based on the two-dimensional image through the timing analysis-optimal solution analysis processing method.

[0021] 6. The optimal solution analysis and processing method is used to optimize the hyperparameters of the timing analysis model, and the optimal hyperparameters (i.e., the number of model units and the Dropout ratio) are obtained. These hyperparameters will be used as input to rebuild and train the timing analysis model, so as to obtain measurement results at high spatial resolution and improve the accuracy of the measurement.

[0022] 7. In the time series analysis model of the processing stage, the time series analysis model will predict the new data according to the rules learned during the training process, remove the noise in the original signal, add the Dropout layer to prevent overfitting, and regularize the model through the noise in the training data. Use the mean square error (MSE) as the loss function, and continuously adjust the model parameters through the training process.

[0023] in, , n is the total number of samples, is the true value of the i-th sample, is the predicted value of the i-th sample.

[0024] 8. In stage 6, the optimal solution analysis and processing method is used to tune the hyperparameters in the time series analysis model. A strategy combining global and local search is adopted to automatically adjust the hyperparameters in the time series analysis model. Through multiple iterations, the performance of each parameter combination is evaluated, and finally the optimal hyperparameter configuration is found to minimize the model loss on the training set.

[0025] 9. Finally, the strain information along the sensing fiber is analyzed by determining the appropriate spectral shift between the reference signal and the measured signal through cross-correlation analysis.

[0026] like Figure 1As shown, a distributed optical fiber sensing system device based on a timing analysis-optimal solution analysis and processing method comprises: a continuous laser output of a tunable laser source is divided into two parts by a coupler 1 (10 / 90 optical coupler), 10% of which is incident on an unbalanced Mach-Zehnder trigger interferometer to provide a trigger signal for an acquisition card, and the rest of the light enters a coupler 2; then the coupler 2 (1 / 99 optical coupler) is divided into two parts, 1% of which is adjusted by a polarization controller 1 so that the "p" and "s" light components have the same power, and 99% enters the sensing optical fiber detection through a circulator and a polarization controller 2, and the sensing optical fiber is a standard single-mode optical fiber; then the interference signal obtained by combining the Rayleigh scattering signal with the 1% laser output from a coupler 3 (50 / 50 optical coupler) is decomposed into "p" and "s" components by a polarization beam splitter; finally, the "p" and "s" lights are collected by the acquisition card.

[0027] like Figure 3 and Figure 4 It can be seen from the comparison that at a spatial resolution of 1.5 mm, the traditional processing method has difficulty in distinguishing different strains and is very likely to cause erroneous analysis, while the method of the present invention can analyze the strain results very well.

[0028] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in the field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for improving the spatial resolution of a distributed optical fiber sensing system, characterized in that: The following steps are involved: S1, collect reference signal and test signal containing strain; S2, perform FFT mapping of the signal to the distance domain and divide it into windows, and calculate the cross-correlation result; S3, reconstructs the 2D image and denoises it using a time series analysis model; S4, uses a global-local hybrid search strategy to optimize model hyperparameters; S5, output high-resolution strain measurement results.

2. The method for improving the spatial resolution of a distributed optical fiber sensing system according to claim 1, characterized in that: Specifically, S1 collects signals twice, one of which is a signal without strain information, which is a reference signal; and the other is a signal containing strain information, which is a test signal.

3. The method for improving the spatial resolution of a distributed optical fiber sensing system according to claim 1, characterized in that: The S2 is specifically: S21, mapping the reference signal and the test signal to the distance domain through fast Fourier transform, and taking a window size A to divide the signal into N equal parts, where the window size A determines the spatial resolution of the system; S22, performing a fast inverse Fourier transform on the local distance domain information of the first reference signal and the measurement signal; S23, performing cross-correlation calculation on the reference signal and the measurement signal after the fast inverse Fourier transform to obtain a one-dimensional cross-correlation result.

4. The method for improving the spatial resolution of a distributed optical fiber sensing system according to claim 1, characterized in that: The S3 is specifically: S31, repeating S22-S23 to obtain the cross-correlation result of each corresponding position of the optical fiber, reconstructing all the obtained one-dimensional cross-correlation results into a two-dimensional image signal at the optical fiber distance, and performing denoising on the image based on the two-dimensional image by using a time series analysis-optimal solution analysis processing method; S32, optimizing the hyperparameters of the time series analysis model by an optimal solution analysis and processing method, and obtaining the optimal hyperparameters, wherein the hyperparameters are the number of model units and the Dropout ratio. These hyperparameters are used as inputs to reconstruct and train the time series analysis model, so as to obtain measurement results at a high spatial resolution and improve the accuracy of the measurement; S33, in the time series analysis model of the processing stage, the time series analysis model will predict the new data according to the rules learned in the training process, remove the noise in the original signal, add the Dropout layer to prevent overfitting, and regularize the model through the noise in the training data; use the mean square error MSE as the loss function, and continuously adjust the model parameters through the training process; in, , n is the total number of samples, is the true value of the i-th sample, is the predicted value of the i-th sample.

5. The method for improving the spatial resolution of a distributed optical fiber sensing system according to claim 1, characterized in that: Specifically, in the S32 stage, the optimal solution analysis and processing method is used to tune the hyperparameters in the timing analysis model, and a strategy combining global and local searches is adopted to automatically adjust the hyperparameters in the timing analysis model. Through multiple iterations, the performance of each parameter combination is evaluated, and finally the optimal hyperparameter configuration is found, thereby minimizing the loss of the model on the training set.

6. The method for improving the spatial resolution of a distributed optical fiber sensing system according to claim 1, characterized in that: Specifically, S5 is to determine the appropriate spectrum shift between the reference signal and the measurement signal through cross-correlation analysis, and analyze the strain information along the sensing optical fiber.

Images