Analysis of laminar and turbulent flows with turbomachinery, biotechnology and biomechanical applications

Sammanfattning: The goal of this study was initially to gain a better understanding of the effects of rotation on turbulent flow in ducts. Knowledge concerning the influence of rotation on the structures of turbulence is of fundamental importance in many applications, e.g. centrifugal separators, turbines or cooling channels in rotating machinery, as well as meteorology and oceanography. Rapidly rotating duct flow is studied experimentally with rotation numbers in the interval [ 0, 1] . To achieve this, in combination with relatively high Reynolds numbers (5000 – 30000 based on the hydraulic radius), water was used as the working medium. The influence of the rotation on the pressure drop in the duct was investigated and suitable scalings of this quantity were studied. Due to questions that arose in the experimental study, two numerical studies were initiated. The first study probed the effect of rotation and geometrical configuration on the development length for turbulent flow, while the second comprised a direct numerical simulation of turbulent flow in a rotating duct. It is shown that while system rotation does not have a marked effect on the development length in a plane channel, the development length is substantially shortened in a duct.Additional systems subject to rotation or curvature effects were studied. The laminar flow of fluid in a rotating PCR-cone was analysed analytically and numerically to understand the increased mixing and temperature homogenization. The flow field in the cone was described and the increased mixing was due to a strong boundary layer flow incited by Coriolis and buoyancy effects. Comparisons of the numerical simulations with experiments yielded good results.A study to quantify the flow of blood in cerebral malformations using three-dimensional videodensitometry was performed. Data from experiments with an idealized flow phantom, as well clinical pathologies, showed that the proposed methodology in conjunction with clinical injection protocols can yield mean flux data with an error less than 20%. Protocol improvements are proposed.