Flow Induced Pulsations Caused by Corrugated Tubes

Abstract

Corrugated tubes can produce a tonal noise when used for gas transport, for instance in the case of flexible risers. The whistling sound is generated by shear layer instability due to the boundary layer separation at each corrugation. This whistling is examined by investigating the frequency, amplitude and onset of the pulsations generated by 2" artificially corrugated tubes and cable feeds. Special attention is given to the influence of the geometry of the corrugations and to the influence of the boundary conditions of the tubes. Two distinct modes are measured. One high mode with a typical Strouhal number Sr = 0.35 and one with a Strouhal number of Sr = 0.1. The relative length scale for the corrugations to be used in the Strouhal number is a modified gap width, which is the gap width excluding the downstream edge radius. The exact Strouhal number for a corrugation is furthermore dependent on details of the corrugation, as the convective velocity of the flow disturbances is influenced by details in the geometry such as edge rounding. The amplitude of the generated pulsations scales with the acoustic pressure (ρcU) and will saturate for higher flow rates (p′/ρcU = constant). The saturation level is independent of pressure and tube length and is solely dependent on the corrugation geometry. Larger cavities will generate higher amplitude pulsations. The onset of the whistling is dependent on the tube itself and the system boundaries. Only for very long tubes is the onset insensitive to the system boundaries and will the onset be determined by corrugations. In that case the onset is determined by a critical boundary layer thickness. For smaller tubes, this critical layer thickness is still relevant, but the boundary conditions will have a large effect. A system with a high reflection coefficient will start at lower gas velocities than a system with a low reflection coefficient. Based on the current results the frequency and amplitude of the pulsations can be predicted.