Modélisation des écoulements diphasiques: Amortissement, forces interfaciales et turbulence diphasique

The first part of the thesis is devoted to two-phase damping. This damping is considered part of the solution against failure of piping, since it constitutes a dominant component of the total damping in piping conveying two-phase flow. However, the mechanisms responsible for two-phase damping are not well understood and no convenient models are available to predict this damping.

Two-phase damping is of the utmost importance to predict fatigue and fretting-wear and consequently predict the life of structures operating in two-phase flow. This damping is also crucial in the prediction of the critical velocity for fluidelastic instability in cross-flow. However, in cross-flow, two-phase damping is accompanied by other significant forces such as two-phase turbulence, viscous damping, quasi periodic forces, etc. A correct measure of two-phase damping is therefore complex. However, in internal flow, especially with clamped-clamped tubes two-phase damping and structural damping are the two only significant damping. This allows a direct measurement of two-phase damping. For this reason, damping experiments were first carried out with internal flow. Nevertheless, the knowledge acquired with internal flow should also be useful to predict two-phase damping in cross-flow as the general mechanism of two-phase damping should be the same. Experiments were performed with a vertical tube clamped at both ends. Two-phase damping ratios were obtained from free transverse vibration measurements on the tube using the log-decrement technique.

A first paper, published in Journal of Fluid and Structures, presents the first conclusions reached. First, gas bubbles of controlled geometry were simulated with glass spheres let to settle in stagnant water. Second, air was injected in stagnant alcohol to generate a uniform and measurable bubble flow.

In both cases, the two-phase damping ratio is correlated to the number of bubbles (or spheres). Two-phase damping is therefore directly related to the interface surface area and, therefore, to flow pattern. Further experiments were carried out on tubes with internal two-phase air-water flow. A strong dependence of two-phase damping on flow parameters in bubbly flow regime was observed. A series of photographs confirms the fact that two-phase damping in bubbly flow increases for a larger number of bubbles, and for smaller bubbles. It is highest immediately prior to the transition from bubbly flow to slug flow regimes. Beyond the transition, damping decreases. It is also shown that two-phase damping increases with the tube diameter.

Part II of the paper proposes a model for pseudo turbulence (bubble induced turbulence) in two-phase flow. Relations for the Reynolds stress tensor for both the dispersed and the continuous phases are proposed depending on Reynolds number Re, void fraction ϵ and viscosity and density ratios (μ, ρ). The proposed relations are useful for all fluid-fluid cases. The implications for bubble size and turbulence forces on bubbles are also studied. A simple model is proposed in this paper as a relation for turbulence forces on bubbles as well as a model to predict bubble size. It is shown that no significant breakup forces are present in pipes conveying two-phase flow. This could explain why the flow pattern transition from bubbly flow to slug flow is independent of tube diameter.

This PhD thesis consists of a study on two-phase damping and influence of void fraction on flow characteristics and on the induced forces. Two-phase damping was identified to be mostly an inertial effect bringing energy to the continuous phase due to a relative motion between phases. This process underlines the major impact of density difference and surface tension. The study on the influence of void fraction on interface force and pseudo turbulence leads to the conclusion of major importance. That only a few percentage of void fraction changes completely the nature of the flow. This shows the danger in identifying single-phase flow phenomena with two-phase flows. In particular, the nature of turbulence in two-phase flow is proved to be completely different from that observed in single-phase flow. (Abstract shortened by UMI.)