浏览全部资源
扫码关注微信
基于神经网络的三阶拉曼放大光传输系统建模
Modeling of Optical Fiber Transmission System with Third Order Raman Amplifier based on Neural Network
- 光通信研究 2025年第5期
- 作者机构：
1.国家电网有限公司信息通信分公司，北京 100761
2.武汉光迅科技股份有限公司，武汉 430205
- 作者简介：
陈佟（1989-），男，江苏南通人。副教授级高工，硕士，主要研究方向为电力通信网及超长距光通信传输技术。
龙函，工程师。E-mail：longhan_91@163.com
- 基金信息：国家电网有限公司科技资助项目(SGZB0000TGJS2301000)
DOI：10.13756/j.gtxyj.2025.240111
中图分类号： TN929收稿：2024-06-05，
修回：2024-06-14，
纸质出版：2025-10-10
- 稿件说明：
移动端阅览
陈佟，邓黎，王谦，等. 基于神经网络的三阶拉曼放大光传输系统建模[J]. 光通信研究，2025(5): 240111.
Chen T, Deng L, Wang Q, et al. Modeling of Optical Fiber Transmission System with Third Order Raman Amplifier based on Neural Network[J]. Study on Optical Communications, 2025(5): 240111.
陈佟，邓黎，王谦，等. 基于神经网络的三阶拉曼放大光传输系统建模[J]. 光通信研究，2025(5): 240111. DOI： 10.13756/j.gtxyj.2025.240111.
Chen T, Deng L, Wang Q, et al. Modeling of Optical Fiber Transmission System with Third Order Raman Amplifier based on Neural Network[J]. Study on Optical Communications, 2025(5): 240111. DOI： 10.13756/j.gtxyj.2025.240111.
摘要
【目的】
2
三阶分布式拉曼放大器（DRA）是目前长距离无中继传输中的光放大前沿技术，与一阶和二阶DRA传输相比，其在传输过程中能实现更优的增益与噪声特性，但传输过程也更为复杂。然而目前对三阶DRA的建模多基于拉曼功率耦合方程，其需要数值方法求解，计算相对复杂。运算量小的建模方法则基于数据驱动型机器学习，需要庞大的数据量，且泛化能力较差。为此需要新的建模方法，在保证模型精度的前提下提升运算速度。
【方法】
2
文章提出了一种基于内嵌物理知识型神经网络（PINN）的方案对三阶拉曼放大光传输系统进行建模。该方案能结合数值方法和数据驱动型机器学习的优点，将微分方程的求解转化为优化问题，通过将微分方程以及边界条件等约束作为损失函数，实现神经网络的训练，使得模型兼顾计算的精度与复杂度。
【结果】
2
文章搭建了一个120 km的C波段47信道的三阶拉曼放大光纤传输系统，并分别采用数值方法和所提的PINN对系统进行功率预测，结果表明，PINN预测的信号功率与传统的数值方法偏差小于0.19 dB，预测的泵浦光功率则几乎无误差，而运算次数则降低了一个数量级。
【结论】
2
PINN能准确预测三阶拉曼放大光传输系统中泵浦光与信号光的功率演化，且相比数值方法显著降低了计算复杂度。
Abstract
【Objective】
2
Third order Distributed Raman Amplifier (DRA) is a novel technology in unrepeatered transmission. Compared with first- and second-order DRA
they offer superior gain and noise performance
yet their propagation dynamics are markedly more intricate. However
the current modeling of third order DRA is mostly based on the Raman power coupling equations
which demand numerical solvers and entail heavy computational loads. Conversely
lightweight data-driven machine-learning models require vast datasets and suffer from limited generalization. New modeling methods are therefore needed to improve computational speed while ensuring the accuracy.
【Methods】
2
Physic Informed Neural Network (PINN) isproposed to model third order Raman amplified optical transmission systems. This method combines the advantages of numerical methods and data-driven machine learning to transform the solution of differential equations into optimization problems. By using constraints such as differential equations and boundary conditions as loss functions
neural network training is achieved
allowing the model to balance the computational accuracy and complexity.
【Results】
2
An 120 km fiber optics transmission system with 47 channels in C-band and amplified by a third order Raman amplifier is built in the paper. Compared with the numerical method
the error of PINN is less than 0.19 dB for the signals
and the results for the pumps are almost the same. Simultaneously
PINN reduced the number of computational operations by one order of magnitude.
【Conclusion】
2
PINN can accurately predict the power evolution of pumps and signals in third order Raman amplified optical transmission systems
and significantly reduce the complexity compared to traditional numerical methods.
关键词
Keywords
references
高建新 , 陈奉洁 , 张新桥 , 等 . 超长距光路子系统拉曼技术应用研究 [J ] . 数字通信世界 , 2023 ( 3 ): 48 - 50 .
Gao J X , Chen F J , Zhang X Q , et al . Application of Raman Technology in Ultra-Long Distance Optical Path Subsystem [J ] . Digital Communication World , 2023 ( 3 ): 48 - 50 .
巩稼民 , 张玉蓉 , 徐军华 , 等 . 二阶拉曼光纤放大器增益特性研究 [J ] . 光子学报 , 2020 , 49 ( 8 ): 0806001 .
Gong J M , Zhang Y R , Xu J H , et al . Research on Gain Characteristics of Second-Order Raman Fiber Amplifier [J ] . Acta Photonica Sinica , 2020 , 49 ( 8 ): 0806001 .
迟荣华 , 吕涛 , 王飞 , 等 . 基于二阶拉曼放大器的OTN超长距光传输系统的理论和实验研究 [J ] . 电子器件 , 2019 , 42 ( 4 ): 887 - 890 .
Chi R H , Lü T , Wang F , et al . Theoretical and Experimental Study on OTN Ultra-Long Distance Optical Transmission System based on Second-Order Raman Amplifier [J ] . Chinese Journal of Electron Devices , 2019 , 42 ( 4 ): 887 - 890 .
汪大洋 , 曹晶 , 李程 , 等 . 二阶拉曼放大器技术在电力超长站距光传输系统中的应用 [J ] . 电力信息与通信技术 , 2017 , 15 ( 8 ): 60 - 65 .
Wang D Y , Cao J , Li C , et al . Application of Second-order Raman Amplifier in Electric Power Ultra-Long Distance Optical Transmission System [J ] . Electric Power Information and Communication Technology , 2017 , 15 ( 8 ): 60 - 65 .
忻向军 , 余重秀 , 任建华 , 等 . 双向抽运拉曼光纤放大器性能的研究 [J ] . 光学学报 , 2004 , 24 ( 6 ): 825 - 829 .
Xin X J , Yu C X , Ren J H , et al . Performance of Bidirectionally Pumped Raman Fiber Amplifier [J ] . Acta Optica Sinica , 2004 , 24 ( 6 ): 825 - 829 .
Zhou J , Chen J , Li X , et al . Robust, Compact, and Flexible Neural Model for a Fiber Raman Amplifier [J ] . Journal of Lightwave Technology , 2006 , 24 ( 6 ): 2362 - 2367 .
Jiang X , Wang D , Chen X , et al . Physics-informed Neural Network for Optical Fiber Parameter Estimation from the Nonlinear Schrödinger Equation [J ] . Journal of Lightwave Technology , 2022 , 40 ( 21 ): 7095 - 7105 .
Zang Y , Yu Z , Xu K , etal . Principle-Driven Fiber Transmission Model based on PINN Neural Network [J ] . Journal of Lightwave Technology , 2022 , 40 ( 2 ): 404 - 414 .
Zang Y , Yu Z , Xu K , et al . Universal Fiber Models based on PINN Neural Network [C ] // 2020 Asia Communications and Photonics Conference (ACP) and International Conference on Information Photonics and Optical Communications (IPOC) . Beijing, China : IEEE , 2020 : 9365071 .
穆宽林 , 武岳 , 周健 , 等 . 基于神经网络和方程数值解的FRA增益控制方法 [J ] . 光通信研究 , 2024 ( 5 ): 230051 .
Mu K L , Wu Y , Zhou J , et al . Gain Control Method for FRA based on Neural Network and Numerical Solution of Equations [J ] . Study on Optical Communications , 2024 ( 5 ): 230051 .
Marcon G , Galtarossa A , Palmieri L , et al . Model-Aware Deep Learning Method for Raman Amplification in Few-Mode Fibers [J ] . Journal of Lightwave Technology , 2021 , 39 ( 5 ): 1371 - 1380 .
Yankov M P , Da Ros F , de Moura U C , et al . Flexible Raman Amplifier Optimization based on Machine Learning-Aided Physical Stimulated Raman Scattering Model [J ] . Journal of Lightwave Technology , 2023 , 41 ( 2 ): 508 - 514 .
0
浏览量
87
下载量
0
CSCD
0
CNKI被引量
- 网络PDF下载
- Meta-XML下载
- BibTeX下载
- Endnote下载
- RefMan 下载
- NoteExpress下载
- NoteFirst下载
关联资源
相关文章
相关作者
相关机构
- 地址：武汉市江夏区高新四路6号烽火科技园4号楼3层 邮编：430205
- 技术支持由北京北大方正电子有限公司提供 鄂ICP备2022010197
鄂公网安备 42011202002092号 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies
- 总访问量：304943 今日访问量：552