Skip to main content
To KTH's start page To KTH's start page

Experimental studies of large particles in Newtonian and non-Newtonian fluids

Time: Thu 2019-09-26 10.15

Location: Kollegiesalen, Brinellvägen 8, Stockholm (English)

Subject area: Engineering Mechanics

Doctoral student: Sagar Zade , Mekanik, SeRC - Swedish e-Science Research Centre, Linné Flow Center, FLOW

Opponent: Professor Ellen Longmire, Department of Aerospace Engineering and Mechanics, University of Minnesota

Supervisor: Luca Brandt, Linné Flow Center, FLOW, Mekanik, Stabilitet, Transition, Kontroll, SeRC - Swedish e-Science Research Centre, Fysiokemisk strömningsmekanik, Processteknisk strömningsmekanik; Fredrik Lundell, Mekanik, Linné Flow Center, FLOW, Wallenberg Wood Science Center

Export to calendar


In everyday human life, laminar flow is arguably an exception whereas turbulent flow is the norm. Yet, the former has been much better understood, naturally since laminar flow renders itself to treatment in a relatively easier fashion compared to turbulence with its chaotic dynamics across multiple scales in space and time. A parallel analogy in terms of sophistication of dynamics can be drawn between single phase and multiphase flows; the latter being the norm yet poorly understood due to numerous complexities arising on account of the huge parameter space involved. It is also remarkable that numerical studies are more prevalent in this field and there is a dearth of experimental results, which are important for both validation purposes and as a beacon to navigate research in practically relevant directions. This work has emerged to address the above issues. The attention has been largely directed towards understanding the flow of spherical particles in a square duct at moderately high concentrations using Particle Image Velocimetry (PIV) with refractive-index-matched (RIM) hydrogel particles. Fluids with Newtonian, viscoelastic and elastoviscoplastic rheology have been investigated due to their presence in natural and industrially relevant flows. Experiments and Direct Numerical Simulations (DNS) with spherical particles in a round pipe with turbulent flow of a Newtonian fluid are also conducted to extend and generalise the observations made in the square duct.

With the ability to optically interrogate the bulk of the flow at high particle concentrations (20\% in this work), many interesting measurements are made possible, focussing on the turbulent regime. For the Newtonian fluid, the pressure drop or, equivalently, the energy required to pump the fluid-particle mixture is a complex function of particle size and concentration in the duct. This phenomenon arises due to the particle concentration distribution, with a local maxima at the core and the walls, and its resulting effect on the dominant stresses in the system i.e.\ the Reynolds shear stress and particle-induced stress. Particles also migrate in a similar fashion in a turbulent flow of viscoelastic suspending fluid but, with a larger tendency to accumulate in the core compared to its Newtonian counterpart at the same Reynolds number leading to a faster rise in total stress with concentration. Finally, for the \textit{thick} elastoviscoplastic fluid, the single-phase flow is laminar but it exhibits turbulence-like fluctuations when particles are added, which are distributed in exotic configurations depending on the interplay between the viscoelastic forces and the ensuing secondary flows as well as inertial forces. On the other hand, a quantitative comparison between simulations and experiments for particles transported along the floor of the duct under turbulent conditions has helped in reinforcing confidence in both approaches.

We believe that these results will establish more confidence in the experimental usage of hydrogel particles for studying the flow of moderately dense suspensions. A natural extension would be the investigation of flow geometries more complex than a pipe or a square duct. Our results at higher Reynolds numbers is expected to motivate numerical simulations which are capable of investigating the detailed causes behind these observations, which are still unclear as of now. The information provided about the overall drag and the associated particle concentration and stress distribution will be helpful in painting a unified picture of turbulent suspension dynamics for a comprehensive range of flow rates and particle sizes. Future studies, either experimental or numerical, bearing similarities or deviations from our observations would also be constructive, for e.g.\ in assessing the sensitivity of the system to parameters that may be overlooked in the present study.