An assessment of rough surface boundary-layer calculation methods

The paper examines typical contemporary predictive methods for rough-wall boundary layer flows, assesses the applicability of various methods, and suggests where improvements might be best made. A typical correlation, an effective wall location differential method, and a discrete roughness element differential method are compared using both compressible and incompressible data sets. Major conclusions are that (1) for sand-grain roughness the expression of Seidman for Stanton number works well if accurate values of skin friction coefficient are used; (2) both differential-based approaches yield acceptable skin friction coefficients when a satisfactory sand-grain equivalent is used; and (3) the algebraic correlations examined give widely varying results for skin friction coefficients, depending on the particular flow conditions.