Turbulent multiphase Taylor-Couette flows

**Luuk Blaauw is a PhD student in the Department Physics of Fluids. (Co)Promotors are prof.dr. D. Lohse and dr. S.G. Huisman of the Faculty of Science & Technology.**

Turbulent multiphase flows are omnipresent in nature and industry. These flows can be divided into flows containing bubbles, droplets, or particles. In this thesis we have studied bubbly flows (chapter 1–2) and particle laden flows (chapter 3) in turbulent Taylor–Couette flow. Chapter 4 focusses on transitional single phase turbulent Taylor–Couette flow. In chapter 1 and 2 we study drag reduction by bubbly flows in salt water in the light of bubbly drag reduction for the shipping industry. Chapter 1focusses on the effect of NaCl, the most common found salt in seawater, on the bubbles and their drag reduction. NaCl inhibits bubble coalescence, leading to a smaller equilibrium bubble size in the flow. These smaller bubbles are less deformable by the flow and have a lower Weber number, making them less effective for bubbly drag reduction. Whereas for fresh water and an air volume fraction of α = 4% we observe a drag reduction of 40%, the drag reduction for salt concentrations around the total salt concentration in seawater is only approximately 15%. We observed a critical salt concentration after which increasing the salt concentration does not affect the drag reduction further.

Chapter 2 extends on the work in chapter 1 by investigating the effects of MgCl2, Na2SO4, and substitute sea salt on the bubbly drag reduction. These salts also inhibit bubble coalescence, leading to smaller bubbles in the flow and less effective bubbly drag reduction. We compare the bubbly drag reductions found for different salt solutions and observe a reasonable collapse of the drag reduction with the ionic strength of the solution, where for solutions with an ionic strength I > 0.7 we observe almost no drag reduction. We separately investigated the effects of NaCH3COO, a salt that does not affect bubble coalescence, on bubbly drag reduction. We observe an increase of the drag reduction with the NaCH3COO concentration. We connect the observed bubbly drag reduction for all salt solutions with the bubble Weber number and find that the bubble Weber number is of utmost importance for effective bubbly drag reduction. The work in chapter 1 and chapter 2 provide new challenges for ship builders implementing drag reducing techniques.

In chapter 3 we focus on turbulent particle-laden Taylor–Couette flow. We use growing hydrogel particles to study particle-laden flows with particle volume fraction from φ ≈ 2% to φ ≈ 75% in one single experiment. This method enables us to study turbulent flows with particle volume fractions not investigated before. For volume fractions φ ≤ 55%, the torque on the inner cylinder increases due to the increasing effective viscosity of the suspension. Higher volume fractions lead to weakening of the Taylor vortices and hence a decrease in the torque. For volume fractions φ ≥ 64% we switch from a volume imposed system to a pressure imposed system, further increasing the torque. At a volume fraction of φ ≈ 75% the internal pressure between particles is so high, they can not move with respect to each other anymore, i.e. the system is completely jammed, and the particles undergo solid body rotation.

In the final chapter we investigate single phase turbulent Taylor–Couette flow at the transition from the classical turbulent regime to the ultimate turbulent regime, i.e. where the boundary layer transitions from laminar to turbulent. By heating and cooling the flow, we increase and decrease the viscosity of the fluid and with that the Taylor number of the flow. The system starts in the classical regime and transitions to the ultimate regime around Ta ≈ 3 × 108. This transition is apparent in both the torque on the inner cylinder where we observe a sudden increase as well as in the velocity at midheight where we also observe a sudden increase. We do not observe a transition back to the classical state when decreasing the Taylor number, characterising the hysteretic nature of the transition. We observe the Taylor vortices before and after the transition and see that there are 12 Taylor vortices in the system before the transition and 10 Taylor vortices in the system after the transition. It is surprising we see an increase in torque when we observe a decrease in the number of Taylor vortices, which are very effective flow structures for the transport of angular velocity. We hypothesise that by transitioning to turbulent boundary layers the Taylor vortices become more turbulent and more effective for the transport of angular velocity, allowing for a higher torque even with less Taylor vortices.