Answers are grounded in this paper and cite their sources.
An adaptive mesh procedure controlled by an error estimate for the Navier–Stokes equations is developed. The mesh can be refined but also coarsened by the mean of an agglomeration algorithm. The error estimation is based on an equation for the discretization error with a source term approximated by the use of a higher order discretized operator. The whole procedure is applied to a turbulent flow around a square-cross section cylinder. The efficiency of the method, evaluated in terms of CPU time and number of cells, shows interesting gains compared to single mesh computations.
No takes yet. Share an insight, caveat, or question.
Hay et al. (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: