This paper presents the results of an extensive experimental investigation of the in-line and transverse forces acting on sand-roughened circular cylinders placed in oscillatory flow at Reynolds numbers up to 1,500,000; Keulegan-Carpenter numbers up to 100; and relative roughnesses from 1/800 to 1/50. The drag and inertia coefficients have been determined through the use of the Fourier analysis and the least squares method. The transverse force (lift) has been analysed in terms of its maximum and root-mean-square values. In addition, the frequency of vortex shedding and the Strouhal number have been determined. The results have shown that all of the coefficients cited above are functions of the Reynolds number, Keulegan-Carpenter number, and the relative roughness height. The results have also shown that the effect of roughness is quite profound and that the drag coefficients obtained from tests in steady flow are not applicable to harmonic flows even when the fluid loading is predominantly drag.
Of the scores of papers dealing with fluid loading on offshore structures none seems to have treated the effect of roughness on the force-transfer coefficients. Yet it is a fact that the structures in the marine environment become gradually covered with rigid as well as soft excrescences. (see Fig. 1). Thus, the fluid loading due to identical ambient flow conditions may be significantly different from that experienced when the structure was clean partly because of the 'roughness effect' of the excrescences on the flow and partly because of the increase of the 'effective diameter' of the elements of the structure.
In the absence of any data appropriate to the harmonic or wavy flows. it has been assumed that "the drag coefficients obtained from tests in steady flow" over artificially - or marine-roughened cylinders "are applicable to wave flows at least when the loading is predominantly drag". Even for large amplitudes of oscillations. there is only a finite vortex street comprised of vortices of nearly equal As the flow reverses, the situation is not that of a uniform flow (with or without free-stream turbulence) approaching a roughened cylinder but rather that of a finite vortex street approaching a rough-walled cylinder. Such a flow cannot be regarded identical to steady flow with some turbulence of fairly uniform intensity and scale as the present results show.
It is a well-known fact that organized, uniform roughness in steady flow about a cylinder precipitates the occurrence of the critical regime and gives rise to a minimum drag coefficient which is larger than that obtained with a smooth cylinder. This is partly because of the transition to turbulence of the free shear layers at relatively lower Reynolds numbers due to disturbances brought about by the roughness elements and partly because of the retardation of the boundary-layer flow by roughness (higher skin friction) and, hence, earlier separation.
