Reduced order modelling of an unstructured mesh air pollution model and application in 2D/3D urban street canyons

A novel reduced order model (ROM) based on proper orthogonal decomposition (POD) has been developed for a finite-element (FE) adaptive mesh air pollution model. A quadratic expansion of the non-linear terms is employed to ensure the method remained efficient. This is the first time such an approach...

Full description

Saved in:
Bibliographic Details
Published in:Atmospheric environment (1994) 2014-10, Vol.96, p.96-106
Main Authors: Fang, F., Zhang, T., Pavlidis, D., Pain, C.C., Buchan, A.G., Navon, I.M.
Format: Article
Language:English
Subjects:
Accuracy
Air pollution
Analysis methods
Applied sciences
Atmospheric pollution
Computer simulation
Exact sciences and technology
Finite element
Finite element method
Mathematical models
Pollution
Proper orthogonal decomposition
Reduced order modelling
Reduced order models
Street canyons
Three dimensional
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A novel reduced order model (ROM) based on proper orthogonal decomposition (POD) has been developed for a finite-element (FE) adaptive mesh air pollution model. A quadratic expansion of the non-linear terms is employed to ensure the method remained efficient. This is the first time such an approach has been applied to air pollution LES turbulent simulation through three dimensional landscapes. The novelty of this work also includes POD's application within a FE-LES turbulence model that uses adaptive resolution. The accuracy of the reduced order model is assessed and validated for a range of 2D and 3D urban street canyon flow problems. By comparing the POD solutions against the fine detail solutions obtained from the full FE model it is shown that the accuracy is maintained, where fine details of the air flows are captured, whilst the computational requirements are reduced. In the examples presented below the size of the reduced order models is reduced by factors up to 2400 in comparison to the full FE model while the CPU time is reduced by up to 98% of that required by the full model. •A new POD reduced order model capable of resolving complex turbulent flows.•POD combined with a LES adaptive mesh model based on an adapted Smagorinsky model.•Efficient treatments of the nonlinear operators via the quadratic expansion method.•Application to 2D/3D turbulent flows and pollution dispersion in urban landscapes.
ISSN:1352-2310
1873-2844
DOI:10.1016/j.atmosenv.2014.07.021