Comparisons of turbulent-flow calculations with experiment

Predictions of the (k, epsilon) turbulence model are compared with experiment for a number of standard cases: fully developed flows in ducts and pipes, developing pipe flow, flows through axisymmetric and rectangular sudden expansions, flows over steps, and axisymmetric swirling flows. In a general way, it appears that the (k, epsilon) model gives good predictions of velocity profiles and good enough estimates of pressure drops to be useful for many engineering applications. Estimates of turbulence intensities and stresses are much less reliable and range from excellent for fully developed flows to unsatisfactory for strongly swirling flows.