Measurements were conducted in the fully developed turbulent flow in a pipe with internal diameter D at a Reynolds number of Re _D= 1.6 × 10 ⁵ . The pipe walls were equipped with regularly spaced square ribs of relative height h/ D= 0.154 , while the pitch-to-roughness height was varied between p/ h= 1.67 and p/ h= 6.67 . The measurements include mean velocity components, Reynolds shear and normal stresses and pressure losses. It is investigated whether the effects of the large roughness on the (time and axially averaged) velocity profile can be described by the classical rough-wall formulation by allowing the value of the von Kármán constant to deviate from its standard value of 0.41.

@en

This thesis describes the results of an experimental study of turbulent flow through a ribbed pipe. The ribbed pipe used is a pipe segment of 50 mm inner diameter with ribs inserted at various pitches (15 to 50 mm). Rectangular and rounded ribs are used, with a rib height of 7.5 or 11 mm. Pressure measurements result in friction factors showing a strong dependency on rib shape and pitch, and for the rounded rib shape also a dependency on Reynolds number. Dynamic pressure measurements show the presence of a frequency that coincides with second harmonic of vortex shedding of the ribs. Measurements with single ribs, rather than ribbed segments, show that the friction behaviour of a ribbed segment cannot be predicted based on a single rib, as the observed friction behaviour scales differently. With PIV measurements, mean flow patterns, shear strength and Reynolds Stress distributions are obtained, confirming the significant impact of rib pitch and shape on the behaviour of the flow that was indirectly observed during the pressure measurements. For the rectangular shaped ribs, axially averaging the results over one pitch flow length results in flow and stress profiles. These profiles are used as input in the theoretical description of flow over a rough wall. With some adjustment, specifically fitting the Von Kármán constant instead of assuming a value a-priori, the results fit within the theoretical description. The friction factors derived in this way, agrees well with the friction factors based on the pressure measurements up to a pitch of 35 mm. A statistical analysis of the flow dynamics shows that the large contributions to the stresses are caused by relatively rare but strong fluctuations, which are, depending on the location in the flow, either ejections or sweeps. The observation of these fluctuations and their location in the flow match with the second harmonic vortex shedding that was hypothesised based on the frequencies in the dynamic pressure spectrum.@en