基于改进 DNN的 F‐P滤波器温度误差补偿方法
CSTR:
作者:
作者单位:

(1.国家电投集团河北电力有限公司,石家庄 050000;2.东方绿色能源(河北)有限公司,石家庄 050000;3.河北省新能源场群智慧控制重点实验室,石家庄 050000;4.华北电力大学电子与通信工程系,河北保定 071003)

作者简介:

肖蔚然(1972-),女,山东省淄博市人,高级工程师,主要从事风电、光伏等新能源发电技术的研究工作;张荣达(1989-),男,工程师,主要从事电力自动化方面的研究工作。

通讯作者:

中图分类号:

TN911. 6

基金项目:

东方绿色能源(河北)有限公司科技项目(20241230034).


Temperature-Error Compensation Method Based on an Improved Deep Neural Network for F-P Filters
Author:
Affiliation:

(1. State Power Investment Corporation Hebei Electric Power Co.,Ltd.,Shijiazhuang 050000,CHN;2. Dongfang Green Energy(Hebei)Co.,Ltd.,Shijiazhuang 050000,CHN;3. Hebei Province Key Laboratory of Intelligent Control for New Energy Field Cluster,Shijiazhuang 050000,CHN;4. Dept. of Electronic and Communication Engineering,North China Electric Power University,Baoding 071003,CHN)

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    为解决 FBG传感系统中 F-P滤波器的温度误差问题,文章提出一种基于改进 DNN的 F-P滤波器温度误差补偿方法。该方法在传统 DNN模型的基础上增加了 BN层、Dropout层、注意力机制层和残差连接层,提高了 DNN模型的性能。利用改进后的 DNN模型预测 F-P滤波器在不同温度和电压时的波长漂移,再将不同温度下 F-P滤波器的波长折算到某一固定温度时的波长,从而消除了温度变化带来的误差。通过搭建实验系统,对所提方法进行实验验证,并与 LSTM和 LSTM+CNN误差补偿模型预测结果进行对比。实验结果表明,误差补偿后的波长标准差最大为 4.7 pm,偏离平均值的最大波动为 13.3 pm,相比补偿前,标准差降低了 99. 96%,最大波动降低了 99. 93%。利用所提方法可在不添加额外硬件的基础上显著提高 FBG传感的精度,减小系统的复杂性和成本。

    Abstract:

    To mitigate the temperature errors in F-P filter in fiber-Bragg-grating(FBG) sensing system,a temperature-error compensation method based on an improved deep neural network(DNN)is proposed for F-P filters. The performance of the DNN model is improved by incorporating a BN layer,a dropout layer,an attention mechanism layer,and a residual connection layer,building on foundation of traditional DNN models. Temperature errors are eliminated using the improved DNN model that predicts the wavelength drift of F-P filters at different temperatures and voltages, and then converts the wavelength of the F-P filters at different temperatures to that at a fixed temperature. An experimental system is developed to experimentally verify the proposed method,and the corresponding results are compared with the prediction results of LSTM and LSTM+CNN error compensation models. The experimental results show that the maximum standard deviation of the wavelength after error compensation is 4. 7 pm,and the maximum fluctuation from the average is13. 3 pm. Compared with before compensation. The standard deviation and fluctuation phase after compensation are 99. 96% and 99. 93% lower than those recorded before applying the compensation method. Therefore,the proposed method can be used to significantly improve the accuracy of FBG sensing,without adding additional hardware,while reducing the complexity and cost of the system.

    参考文献
    相似文献
    引证文献
引用本文

肖蔚然,韩晓冉,张荣达,刘卜玮,刘鹏鲲,姚国珍.基于改进 DNN的 F‐P滤波器温度误差补偿方法[J].半导体光电,2026,47(1):186-192. XIAO Weiran, HAN Xiaoran, ZHANG Rongda, LIU Bowei, LIU Pengkun, YAO Guozhen. Temperature-Error Compensation Method Based on an Improved Deep Neural Network for F-P Filters[J].,2026,47(1):186-192.

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2025-10-13
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2026-03-20
  • 出版日期:
文章二维码