CFD offers a useful tool for natural ventilation studies. Its use is increasing as algorithms become more sophisticated and computers become more powerful. For more innovative designs its capabilities are particularly useful as standard design methods and codes may not be applicable. However, the use of CFD is not straightforward and there are significant difficulties that users, particularly novice users, face. In all applications care must be given to the selection of appropriate geometries, meshes and models. Additionally, verification and validation must be carried out.
One crucial aspect of model selection is whether to use a steady state or time dependent model. Steady state has the advantage of shorter computer times and of not having to address the questions of temporal accuracy in the solution procedure. Additionally, the results are easier to analyse as there is only one set of variables calculated at each spatial point. However, a steady state solution is deficient in several regards: it does not capture the effects of coherent flow structures such as large eddies and RANS models often impose symmetry on the solution that is unrealistic. The latter leads to the phenomenon of no cross-stream ventilation through symmetric openings on a symmetric building, when the wind direction is at 90o to the face of the building without openings. Some researchers have used unsteady RANS, but this does not capture all the unsteadiness in the flow and there are questions as to how to reconcile an unsteady solution of a time-averaged turbulence model.
The technique that offers a more consistent way of incorporating unsteady effects is Large Eddy Simulation and its more recent derivative, Detached Eddy Simulation. The latter combines RANS and LES in a consistent manner that allows for LES to be used when the grid size is sufficiently fine and reverts to RANS when this is not the case.
In this paper results are presented for a symmetric cube with identical openings on opposite faces. Results from DES are compared with earlier results from RANS and with full-scale data collected at the Silsoe Research Institute.
英国ノッティンガム大学･環境流体力学准教授。1984年にブリストル大学数学科を卒業後、1988年にリーズ大学で機械工学博士号を取得。ブラッドフォード大学、リーズ大学、シティ大学を経て、1993年にノッティンガム大学の講師となる。その後、2000年に上級講師、2003年に准教授。現在、ノッティンガム基幹施設センター長と大学内CFDグループの議長を務める。
ナイジェル・ライト博士は、流体力学分野での数値計算手法の開発・応用の経験を、 20年以上にわたり有する。環境流体問題への数値計算の応用に関心を持ち、川及び河口近傍流れのモデル化や、風工学、汚染物質拡散を一例とした多様な分野で活動している。現在の研究内容は、建物への風の影響、自動車の横風影響、建物の構造と流体の相互作用をモデル化するための数値流体力学(CFD)と有限要素解析(FEA)の連成手法。(翻訳：西澤)
Copyright (C) 2005 BRI All Rights Reserved.