[PhD Defence] Yash Kulkarni скачать в хорошем качестве

[PhD Defence] Yash Kulkarni

Theoretical and numerical modelling of the three-phase dynamic contact line and atomisation Jury : Jens Eggers, Professeur, Université de Bristol — Rapporteur Shervin Bagheri, Professeur, KTH Royal Institute of Technology — Rapporteur David Quéré, Directeur de recherches, CNRS — Examinateur Lydia Bourouiba, Professeur, Massachusetts Institute of Technology — Examinatrice Sébastien Tanguy, Maître de conférences, Université Paul Sabatier — Examinateur Stéphane Zaleski, Professeur des Universités, Sorbonne Université — Directeur de thèse Stéphane Popinet, Directeur de recherches, CNRS — Directeur de thèse Abstract : The three-phase moving contact line is known to have a force singularity at the triple point of the contact line in the traditional no-slip boundary condition. A remedy to this singularity is to allow fluid to have a small slip at the fluid-solid surface. This Navier slip boundary condition is known to resolve the force singularity. In part 1 of the thesis, we begin by analytically showing the smoothness properties of three slip-like boundary conditions, the Navier slip (NBC), the super slip and the generalised Navier boundary condition (GNBC). We show that the velocity flow field for the Navier slip is C0, for the super-slip is C1 and for the GNBC is C2. We then use the Navier slip boundary condition to investigate the dynamic wetting failure in a rapidly advancing contact line. A rapidly advancing contact line can be obtained by letting a feed flow of velocity V fall from a height H on a plate moving horizontally at velocity U. For given fluids and a given geometry the problem has two control parameters, a Reynolds Re number based on V and a capillary number Ca based on U. Beyond a critical condition wetting failure happens and no steady-state solutions are found. This "curtain coating" or "hydrodynamic assist" setup is extremely challenging to simulate in the conditions of the experiments because of the centimetre-to-nanometer length scale ratio and an advancing Ca of order 1. We show that with a constant contact angle and the NBC, at the limit of our resolution, we obtained the stability window as a function of slip length. The anomalous accelerating flow field in the vicinity of the contact line could be explained by inertia. We also numerically developed an inertially corrected Stokes flow wedge solution (IC-SFW) that predicts the velocity field at the micrometer scale. We then develop a free angle methodology to impose a flow-consistent contact angle allowing us to implement the GNBC in the sharp interface limit. This GNBC is tested on the pulling plate setup and is shown to fully regularise the singularity at the triple point. Finally, to test the intermediate region solutions in the contact line, we numerically design a large nanopore setup. This setup allows us to narrow down the ratio of largest to smallest length scales, allowing us to fully probe the flow with high-resolution simulation. In part two of the thesis, we do direct numerical simulation (DNS) of atomisation. We first study the inertial rupture of a sessile droplet subject to impulsive acceleration. The results are compared with the experiments and a tracer study is done having implications for pathogen transmission. Finally, we attempt the problem of atomising pulsed jet at the limit of our resolution. We show that the default Volume of Fluid (VoF) method gives grid-dependent rupture causing grid-dependent statistics on the drop size distribution. To remedy this we use the Manifold Death (MD) method that allows us to perforate thin sheets at a critical thickness hc larger than the grid size. We show statistically converged drop size distribution can be obtained with MD-VoF above the critical thickness hc. The thesis also contains two short notes on vibrating drops and a fractal-like behaviour of the atomising interface.