基于小波字典的风速数据重构的压缩感知方法
A Wavelet Dictionary-Based Compressive Sensing Method for Reconstruction of Wind Speed Data
朱一凯 1余哲帆 1陈安妮 1万华平1
作者信息
- 1. 东南大学混凝土及预应力混凝土结构教育部重点实验室 南京,211189;浙江大学建筑工程学院 杭州,310058
- 折叠
摘要
由于风速具有明显的非平稳性,常用字典的压缩感知(compressed sensing,简称CS)方法对于风速信号重构效果不佳,故引入基于小波字典的压缩感知方法,用于重构风速缺失数据,有效提升了风速信号的重构精度.通过风速仿真数据和广州塔的监测风速数据验证了本研究方法的有效性,并研究了数据缺失工况、正则化参数、小波字典层数和小波类型对风速信号重构效果的影响.结果表明,基于小波字典的压缩感知方法可有效重构缺失的风速信号.
Abstract
Wind speed is usually non-stationary and not naturally sparse.The commonly used dictionary for the compressed sensing(CS)method is not effective for the reconstruction of the wind speed signals.In this paper,the wavelet dictionary is introduced to improve the sparsity of wind speed signals and effectively enhance the ac-curacy of the reconstructed signal.The effectiveness of this method is verified using both wind speed simulation data and monitored wind speed data of Canton Tower.The effects of data missing scenarios,regularization pa-rameters,wavelet dictionary layers,and wavelet types on the reconstruction performance of the CS method are explored in detail.The results show that the wavelet dictionary-based CS method has high accuracy in recon-structing the missing wind speed signals.
关键词
压缩感知/风速/数据重构/小波字典Key words
compressive sensing/wind speed/data reconstruction/wavelet dictionary引用本文复制引用
基金项目
国家重点研发计划(2021YFF0501001)
国家自然科学基金(51878235)
浙江省重点研发计划(2021C03154)
混凝土及预应力混凝土结构教育部重点实验室开放基金(CPCSME2020-05)
出版年
2024
NETL
NSTL
万方数据