The identification of coherent structures using proper orthogonal decomposition and dynamic mode decomposition

本征正交分解 分解 计算机科学 数学 鉴定(生物学) 希尔伯特-黄变换 分解法(排队论) 振动 奇异值分解 算法
作者
Qingshan Zhang,Yingzheng Liu,Shaofei Wang
出处
期刊:Journal of Fluids and Structures [Elsevier BV]
卷期号:49: 53-72 被引量:100
标识
DOI:10.1016/j.jfluidstructs.2014.04.002
摘要

Abstract A comprehensive comparison was conducted on the identification of coherent structures in fluid flow using Proper Orthogonal Decomposition (POD) and Dynamic Mode Decomposition (DMD). The influences that multi-dominant structures and high-order harmonics had on the decomposed modes were taken into extensive consideration. To this end, a series of fabricated patterns was constructed for the benchmark testing to simulate multi-dominant convective structures superimposed in a stationary field. The comparison showed that the temporal DMD method could clearly separate each structure in the spatial and spectral senses, while the POD mode corresponding to the desired structure is contaminated by the other uncorrelated structures. Subsequently, two case studies of the real wake flows, which were determined from high-repetition TR-PIV measurements, were employed to demonstrate the discrepancies of the POD and DMD algorithms in extracting coherent structures. For the wake flow behind a single cylinder at Re D =8000, the temporal DMD algorithm accurately determined the frequency, wavelength and convection speed of the Karman-like vortex street and its higher-order harmonics. However, although the first two POD modes are closely related to the Karman-like vortex street, the higher POD modes embedded as larger structures are obscure in the physical significance. Finally, the wake flow behind two side-by-side cylinders of different diameters at Re D =1000 based on the diameter of the small cylinder was measured; two configurations with different gaps were chosen for comparison, i.e., G/D=0.5 and 2.0. For the wake flow at G/D=0.5, the POD and DMD algorithms determined the major features of the single-dominant structure. For the wake flow at G/D=2.0, the first and second temporal DMD modes effectively and independently extracted the Karman-like vortex structures behind the large and small cylinders, respectively. Meanwhile, although the first and second pairs of POD modes generally captured these two convecting structures, respectively, there was obvious existence of the undesirable contamination of the POD mode, as reflected in the interaction between the desired and uncorrelated structures.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
怡然冰之完成签到,获得积分10
1秒前
圆滚滚完成签到,获得积分10
1秒前
整齐的惮完成签到 ,获得积分10
1秒前
Inevitable完成签到,获得积分10
2秒前
2秒前
眼睛大的仰完成签到,获得积分10
2秒前
2秒前
顾矜应助小郭采纳,获得10
2秒前
guoyanna发布了新的文献求助10
2秒前
2秒前
3秒前
香豆素完成签到 ,获得积分10
3秒前
3秒前
yuyukeke发布了新的文献求助10
3秒前
柔弱修洁关注了科研通微信公众号
3秒前
善学以致用应助0ne222采纳,获得10
3秒前
闪闪落雁完成签到,获得积分10
3秒前
3秒前
4秒前
LSX2823完成签到,获得积分10
5秒前
5秒前
帅气的亦绿完成签到,获得积分20
6秒前
6秒前
小铭同学完成签到,获得积分10
7秒前
7秒前
7秒前
Unicorn发布了新的文献求助30
8秒前
8秒前
fighting发布了新的文献求助10
8秒前
9秒前
迷你的颖发布了新的文献求助10
9秒前
9秒前
Antheali应助JUGG采纳,获得10
9秒前
零下八点半结冰完成签到,获得积分20
10秒前
脑洞疼应助zyf采纳,获得10
10秒前
Feng完成签到,获得积分10
10秒前
小汤发布了新的文献求助10
11秒前
刻苦的三问应助lei采纳,获得10
11秒前
11秒前
Archy完成签到,获得积分10
11秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
Acute Mountain Sickness 2000
Handbook of Milkfat Fractionation Technology and Application, by Kerry E. Kaylegian and Robert C. Lindsay, AOCS Press, 1995 1000
A novel angiographic index for predicting the efficacy of drug-coated balloons in small vessels 500
Textbook of Neonatal Resuscitation ® 500
The Affinity Designer Manual - Version 2: A Step-by-Step Beginner's Guide 500
Affinity Designer Essentials: A Complete Guide to Vector Art: Your Ultimate Handbook for High-Quality Vector Graphics 500
热门求助领域 (近24小时)
化学 医学 生物 材料科学 工程类 有机化学 内科学 生物化学 物理 计算机科学 纳米技术 遗传学 基因 复合材料 化学工程 物理化学 病理 催化作用 免疫学 量子力学
热门帖子
关注 科研通微信公众号,转发送积分 5072099
求助须知:如何正确求助?哪些是违规求助? 4292584
关于积分的说明 13375086
捐赠科研通 4113598
什么是DOI,文献DOI怎么找? 2252529
邀请新用户注册赠送积分活动 1257381
关于科研通互助平台的介绍 1190193