测绘通报 ›› 2022, Vol. 0 ›› Issue (3): 152-156.doi: 10.13474/j.cnki.11-2246.2022.0095

• 技术交流 • 上一篇    下一篇

基于ICEEMDAN的多滤波算法在超高层动态变形监测中的应用

熊春宝1, 庞红星1, 王猛1, 史青法2   

  1. 1. 天津大学建筑工程学院, 天津 300072;
    2. 天津市陆海测绘有限公司, 天津 300191
  • 收稿日期:2021-03-16 修回日期:2021-12-21 出版日期:2022-03-25 发布日期:2022-04-01
  • 作者简介:熊春宝(1964-),男,博士,教授,主要从事结构工程健康监测。E-mail:luhai_tj@126.com
  • 基金资助:
    国家自然科学基金面上项目(51578370)

Application of multi-filter algorithm based on ICEEMDAN in dynamic deformation monitoring of super high-rise structures

XIONG Chunbao1, PANG Hongxing1, WANG Meng1, SHI Qingfa2   

  1. 1. School of Civil Engineering, Tianjin University, Tianjin 300072, China;
    2. Tianjin Surveying and Hydrography Co., Ltd., Tianjin 300191, China
  • Received:2021-03-16 Revised:2021-12-21 Online:2022-03-25 Published:2022-04-01

摘要: 针对监测数据中存在多路径误差和随机噪声的问题,本文提出了一种基于改进的带有自适应噪声的完备集合经验模式分解(ICEEMDAN)、小波包分解(WP),以及递归最小二乘算法(RLS)的联合滤波算法(IWPR)。该算法首先对原始信号进行ICEEMDAN分解,得到一系列本征模态函数(IMF)分量;然后基于标准化模量的累积均值将IMF划分为高频IMF和低频IMF;最后考虑相关系数,利用WP和RLS分别对高频IMF、低频IMF进行去噪,重构两者降噪信号,获得动态位移响应。结果表明:相对于单一算法EMD、CEEMDAN、ICEEMDAN等,IWPR算法能够更有效地消除多路径误差和随机噪声,从而提高超高层GNSS RTK监测数据的精度。

关键词: GNSS RTK;超高层结构;动态变形监测;多滤波;降噪

Abstract: Aiming at the multi-path errors and random noise in the monitoring data,a combined algorithm (IWPR) based on improved complete ensemble empirical mode decomposition with adaptive noise,wavelet packet decomposition and recursive least square algorithm is proposed.Firstly,the original signal is decomposed by ICEEMDAN to obtain a series of intrinsic mode function (IMF) components.Then,IMFs are divided into high-frequency IMF and low-frequency IMF on the basis of the mean of standardized accumulated modes (MSAM).Finally,considering the correlation coefficient,WP and RLS are employed to denoise the high-frequency and low-frequency IMF respectively.In order to acquire the dynamic displacement response of structures,both signals denoised by means of WP and RLS will be reconstructed.The results indicate that,compared with single algorithm EMD,CEEMDAN and ICEEMDAN,IWPR algorithm can eliminate the multi-path errors and random noise more effectively.And this algorithm can effectively improve the monitoring data accuracy of GNSS RTK.

Key words: GNSS RTK;super high-rise structures;dynamic deformation monitoring;multi-filter;denoising

中图分类号: