Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Positioning method for Sagnac distributed optical fiber sensing system based on convolutional neural network ensemble learning

A convolutional neural network and integrated learning technology, applied in the positioning field of distributed optical fiber sensing system, can solve the problems of complex positioning process and loss of distributed optical fiber sensing, etc., and achieve the effect of simple positioning process, noise insensitivity and high efficiency

Active Publication Date: 2021-03-23
SHANGHAI UNIV
View PDF11 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Some scholars use the support vector machine (SVM) regression model to make the measured position close to the real position, but it needs to use the zero frequency method for preliminary positioning in advance, and the positioning process is complicated
The applicant once proposed a discretization positioning method based on a machine learning classification model of a Sagnac distributed optical fiber sensing system, which transformed the disturbance positioning problem into a multi-classification problem of interference signals caused by disturbances in the positions of different sensing fibers. However, this The classification method needs to collect interference signals at all disturbance locations, and loses the advantage of continuous monitoring of distributed optical fiber sensing, which has become an urgent technical problem to be solved

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Positioning method for Sagnac distributed optical fiber sensing system based on convolutional neural network ensemble learning
  • Positioning method for Sagnac distributed optical fiber sensing system based on convolutional neural network ensemble learning
  • Positioning method for Sagnac distributed optical fiber sensing system based on convolutional neural network ensemble learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] In this example, see Figure 1-2 , a method for positioning a Sagnac distributed optical fiber sensing system based on convolutional neural network ensemble learning, comprising the following steps:

[0034] 1) Taking the points at fixed intervals on the sensing fiber as the disturbance points, the sensing system obtains the interference signals generated by simulating disturbances at each point respectively, and after preprocessing, select a part as the training set and the other part as the verification set;

[0035] 2) Train two convolutional neural network CNN models for different loss functions, so that the two CNN models can accurately locate the near-end and far-end disturbances respectively; optimize the parameters through the verification set to obtain the best training effect;

[0036] 3) Combine the training results of the two CNN models through the integrated learning method to obtain the final disturbance position prediction model based on CNN integrated le...

Embodiment 2

[0041] This embodiment is basically the same as Embodiment 1, especially in that:

[0042] In this embodiment, the preprocessing in step 1) and step 4) includes: obtaining the spectrum of the interference signal or the frequency spectrum of the interference signal, performing normalization processing, and determining an appropriate length of spectrum data.

[0043] The integrated learning method in the step 3) adopts Stacking, Bagging or Boosting integrated learning methods.

[0044] This embodiment can ensure the continuous monitoring of the distributed optical fiber sensing system, and can predict the position of any unknown disturbance on the sensing optical fiber, the positioning process is simple, and the efficiency is high.

Embodiment 3

[0046] This embodiment is basically the same as the above-mentioned embodiment, and the special features are:

[0047]In this embodiment, due to the lack of conditions for the time being, it is impossible to actually collect various data required by the disturbance position prediction model, so the OptiSystem software is used to simulate the monitoring of pipeline leakage by the annular Sagnac distributed optical fiber sensing system to verify the volume-based monitoring of this embodiment. Feasibility of a localization method based on ensemble learning of product neural networks.

[0048] like figure 1 As shown, the simulated annular Sagnac distributed optical fiber sensing system includes a continuous laser 1, a 2×2 bidirectional 3dB optical coupler 2, a sensing fiber 3, a sensing fiber 4, a delay fiber 5, a phase modulator 6, and a photodetector 7. Data acquisition and processing unit 8. Length R of sensing fiber 3 and sensing fiber 4 1 , R 2 The sum is equal to the len...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a positioning method for a Sagnac distributed optical fiber sensing system based on convolutional neural network ensemble learning. The method comprises the steps of taking a fixed interval point on a sensing optical fiber as a disturbance point, enabling the sensing system to obtain an interference signal generated by the simulation disturbance of each point, carrying outthe preprocessing of the interference signal, and enabling the preprocessed interference signal to serve as a training set and a verification set, training two convolutional neural network models fordifferent loss functions to enable the two CNN models to accurately position near-end and far-end disturbances respectively, combining training results of the two models through an ensemble learning method to obtain a disturbance position prediction model based on CNN ensemble learning, and optimizing parameters of each model through the verification set. A to-be-positioned interference signal ispre-processed as a test sample, and the to-be-positioned interference signal is tested by using the trained prediction model to obtain a disturbance position of the to-be-positioned interference signal. The method does not need signal demodulation, is low in system complexity, is not sensitive to noise, is simple in data processing method, is stable and accurate in positioning result, and can be used for the disturbance positioning of an annular or linear Sagnac distribution optical fiber sensing system.

Description

technical field [0001] The invention relates to a positioning method for a distributed optical fiber sensing system, in particular to a positioning method for a Sagnac distributed optical fiber sensing system based on convolutional neural network (CNN) integrated learning. Background technique [0002] Distributed optical fiber sensing technology has broad application prospects in the detection and location of high-pressure pipeline leaks. Among them, the Sagnac distributed optical fiber sensing system has the advantages of strong anti-interference and low light source requirements, and is one of the current research hotspots. The zero frequency method is the main positioning method of the Sagnac distributed optical fiber sensing system, but the zero frequency is often submerged in the noise, which affects the positioning accuracy. The accuracy of positioning can be improved by improving the structure of the sensing system and using phase-generated carrier homodyne demodulat...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G01D5/353G06N3/04G06N3/08G06N20/20
CPCG01D5/35322G06N3/08G06N20/20G06N3/045
Inventor 方捻吕继东王陆唐王春华
Owner SHANGHAI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com