基于卷积神经网络的分布式光纤温度传感系统  

Distributed Fiber Optic Temperature Sensing System Based on Convolutional Neural Network

作　　者：温思行 全源[1] WEN Si-xing;QUAN Yuan(Dongguan Power Supply Bureau of Guangdong Power Grid Company,Dongguan 523000 China)

机构地区：[1]广东电网公司东莞供电局,广东东莞523000

出　　处：《自动化技术与应用》2025年第1期150-153,共4页Techniques of Automation and Applications

基　　金：广东省电网科研项目(GJ140002);东莞市级课题:松山湖智慧能源示范专项(031900GS62200156)。

摘　　要：针对地下电缆温度监测问题,提出一种基于卷积神经网络的分布式拉曼温度传感系统。由于拉曼散射光非常弱,因此该系统的信噪比对于系统性能的影响非常大,如何有效地去除噪声目前成了该领域的关键问题。传统去噪方法有的需要对传统传感系统的硬件进行调整,有的消耗时间较长不灵活,都无法较好地且方便地实现去噪功能。提出引用卷积神经网络对大量仿真数据进行学习,然后用仿真数据和真实数据对该训练好的模型进行验证,实验表明,基于卷积神经网络的去噪算法能够有效地去除分布式拉曼温度传感系统的噪声。For the underground cable temperature monitoring problem,this paper proposes a distributed Raman temperature sensing system based on convolutional neural network,because the Raman scattered light is very weak,so the signal-to-noise ratio of this system has a great impact on the system performance,and how to effectively remove the noise is now become a key problem in this field.Some traditional denoising methods need to adjust the hardware of the traditional sensing system,and some consume longer time and are inflexible,all of which cannot achieve the denoising function better and conveniently.In this paper,it proposes to use Convolutional Neural Network to learn a large amount of simulation data,and then validates the trained model with simulation data and real data.Experiments show that the noise removal algorithm based on Convolutional Neural Network can effectively remove the noise of distributed Raman temperature sensing system.

关 键 词：分布式光纤传感 拉曼散射 卷积神经网络 去噪 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP212.11[自动化与计算机技术—控制科学与工程]