Simulation and prediction of non-Newtonian flows

Abstract: In the talk, we will discuss two problems related to (i) simulation of non-Newtonian flow (-- bubble bursting in an elasto-viscoplastic medium) and (ii) predictions in non-Newtonian flows (-- elastic stress prediction in a viscoelastic turbulent channel flow).

Bursting bubbles at a liquid-gas interface is ubiquitous in both natural processes and engineering applications ranging from oceanic wave breaking to food processing. The gas bubble at the free surface can rupture leading to formation of a bubble cavity. The open cavity is unstable due to large surface energy and thereby collapses leading to a sequence of dynamic events including capillary wave propagation, Worthington jet and droplet formation. Apart from inertia and surface tension, the rheological properties of the liquid can have a huge influence on the dynamics of bubble bursting. Using direct numerical simulations, we simulate bubble bursting (the evolution of open cavity) in an elasto-viscoplastic medium to explore the complex interplay between elasticity and yield stress.

The turbulent flow of viscoelastic fluids have piqued the interest in drag-reduction and flow control communities as tiny amount of polymer could alter the near-wall cycle leading to reduction of friction drag in pipe flows. A complete description of viscoelastic turbulence would require characterization of both velocity and polymeric stresses. However, accessing polymer deformation directly from experimental measurements remains a challenging goal. On the other hand, direct numerical simulations are limited by huge computational costs, as polymeric stresses are characterized at even finer scales compared to Newtonian turbulence. Further, neural network models have exhibited excellent capabilities in predicting the instantaneous state of Newtonian turbulent flow. Hence, in the present study the idea of non-intrusive sensing has been applied to viscoelastic turbulent channel flow to predict the velocity fluctuation and polymeric stress components near the wall using quantities measured at the wall.