ASSESSMENT OF ARTIFICIAL DISSIPATION MODELS FOR 3-DIMENSIONAL INCOMPRESSIBLE-FLOW SOLUTIONS

Various approaches for constructing artificial dissipation terms for t hree-dimensional artificial compressibility algorithms are presented a nd evaluated. Two, second-oi der accurate, central-differencing scheme s, with explicitly added scalar and matrix-valued fourth-difference ar tificial dissipation respectively, and a third-order accurate flux-dif ference splitting upwind scheme are implemented in a multigrid time-st epping procedure and applied to calculate laminar flow through a stron gly curved duct. Extensive grid-refinement studies are carried old to investigate the grid sensitivity of each discretization approach. The calculations indicate that even the finest mesh employed, consisting o f over 700,000 grid nodes, is not sufficient to establish grid indepen dent solutions. However, all three schemes appear to converge toward t he same solution as the grid spacing approaches zero. The matrix-value d dissipation scheme introduces the least amount of artificial dissipa tion and should be expected to yield the most accurate solutions on a given mesh. The flux-difference splitting upwind scheme, on the other hand is more dissipative and, thus, particularly sensitive to grid res olution but exhibits the best overall convergence characteristics on g rids with large aspect ratios.