Quantifying acceptable mesh dependencies for computational investigations of airflows within rooms

Abstract

Ignorance of mesh dependency can sometimes be an embarrassment in numerical calculations. Mesh structures need to be developed to eradicate mesh dependency without compromising the finite computing resource and/or incurring large computational expense. For the inexperienced, mesh deployment often degenerates into a time-consuming iterative procedure. To understand mesh influences, different mesh-structures have been applied to two simple, yet typical, room–air–ventilation configurations. The predictions, using various mesh regimes, have been compared with a datum result set: marks have been awarded according to the closeness of comparison to mesh-independent predictions. This mark, i.e. the ‘Mesh Independence Number’ (MIN), is a numerical indicator of the closeness to achieving a mesh in-dependent solution. A relationship for the MIN versus number of nodes is presented together with relationships for the average values of the air-velocity components. In many instances, high correlation-coefficients are established which suggests that this MIN methodology could reduce the need for iterative time-consuming mesh-dependency studies.