DrLyes Kahouadji

My research interests span several areas including: Applied Mathematics; numerical analysis, Fluid-Structure Interaction, adaptive mesh refinement, Miscible and immiscible multiphase flows; heat and mass transfer, and phase change; interfacial flows driven by thermocapillarity and surfactant-driven effects; microfluidics; numerical methods for CFD; high-performance computing; Newtonian and non-Newtonian fluids including viscoplastic materials.

It is very common in nature as in industrial processes that two or more phases are present simultaneously. One of the most striking examples is probably the aerodynamic breakup atomization caused by shear stresses at the liquid-gas interface. This process consists of a liquid jet projected at high speed in a gas that destabilizes gradually until breaking up into a myriad of droplets. This mechanism, which is widely used in fuel injection or perfumery, for example, is only one aspect of a broad class of phenomena that can be gathered under the name of Multiphase Flows. General fluid-fluid interfaces are found in an extraordinary variety of situations and scales such as free-surface waves, jets, bubbles and drops just to name
a few examples. From a numerical perspective, the typical issues faced in multiphase flows lead immediately to questions such as: What is the actual shape of the interface or what are the interactions between the phases?

In the wide field of multiphase flow, analytical approaches are useful but extremely limited to the most trivial cases. The natural extension is Direct Numerical Simulation (DNS), helped by the continuing growth of information technology resources. Such flows are generally three-dimensional and require advanced numerical methods for many reasons. First, physical parameter discontinuities at the interface require particular techniques and several popular methods could be chosen such as Volume of Fluid (VOF) [ 1], Level Set [3], Front Tracking [ 5] or their recently developed hybrid combinations. The history of the development and descriptions of such methods for multiphase flows is thoroughly covered in [ 4]. The discretized systems involve different and widely ranging spatio-temporal scales and therefore a large number of degrees of freedom. It is often necessary to scan the parameter space to study robustness of a system. Our approach that I am using to handle any multiphase configuration is based on an innovative hybridization of the Front-Tracking and Level Set methods which is fully parallelized and able to run on a variety of computer architectures.

The code BLUE: High Performance Computing for Complex Multiphase Flows:

Any industrial, civil or environmental application requires the knowledge of several topics of fluid dynamics (Heat transfer, Fluid-Structure Interactions, Turbulent/Laminar flow, ...). Moreover, it is very rare to have all these knowledge implemented in a single numerical framework. The objective of my research projects is the continued expansion of the frontier of computational simulations of flows through the exploitation of our newly developed code for the simulation of multiphase, multiphysics and incompressible flows. In conjunction with colleagues at LISN-CNRS (France) and Hongik University (Korea), we have developed a robust high performance, parallel code called BLUE which uses high fidelity front-tracking/immersed boundary algorithms for Lagrangian tracking of arbitrarily deformable phase interfaces including breakup and coalescence and a precise treatment of surface tension forces. High performance calculations on over 250k processor threads have been achieved with excellent scalability. This combination of massive parallelization and precision interface treatment allows for the numerical solution of a wide variety of multiphase flow scenarios: free surface instabilities, flow of falling films, bubbles or drops with coalescence and breakup and flow around immersed solid objects for microchannel flows are a few examples. The code has been extended to include thermal and species transport with phase change in order to be able to treat the physics of boiling, solidification or multi-species physics as well as Coriolis (via an analytically exact time-integration method) and centrifugal forces to allow for simulations in rotating systems. Additional physics modules for turbulence have also been developed and recently we have developed modules for the treatment of surfactant laden flows, miscible flows and full fluid-structure interaction. The immersed boundary concept is used to easily define complex solid shapes and in addition stereolithography data files can be input or output in order to handle industrially relevant complex solid geometries. We are generalizing our front tracking/immersed boundary concepts to extend to a wide range of simulations of greater physical complexity (rheology, chemistry, geometry), even higher spatial resolution and larger number of processor threads in order to advance the frontier of what is capable in direct simulation of highly complex 3D multiphase flows in high performance computing. In particular, we plan to expand our capabilities in simulating complex rheology (non-Newtonian fluids) in microfluidic encapsulation processes (drop within a drop), chemistry (surfactant laden flows) and complex solid geometry including full fluidstructure interaction with both rigid and deformable solids (neo-Hookean). All these activities has generated great interest and resulted in many collaborative projects nationally and internationally: U. Birmingham, University College London, Hongik Univ. Korea, Santiago Univ. Chile, PMMH-ESPCI Paris, and City Univ. of Hong Kong.

I have carried out the above work as part of the Matar Fluids Group (at Imperial College London) in close collaboration with multi-national industrial partners in the energy, fast-moving consumer goods, pharmaceuticals and agrochemicals, as well as automation sectors, which include:

○ ABB (turbulent flows past sensors in horizontal pipes, providing assistance for a project led by prof. Daryl Williams);
○ BP (droplet entrainment in turbulent two-phase flows);
○ P&G (all simulation results are validated against P&G experimental and pilot-scale data):
            - Mixing of miscible fluids at high viscosity ratios in stirred vessels;
            - Downwards turbulent annular flows with thermal effects for falling film reactors;
            - Displacement of high-viscosity fluids (Newtonian and viscoplastic) by turbulent lower viscosity fluids in canonical and complex channel geometries for cleaning and decontamination applications;
            - Turbulent aeration of fluids in stirred vessels;
            - Flows of pastes through static mixers;
            - Two- and three-phase flows and encapsulation in microchannels;
            - Control of capillary breakup for bottle-filling applications.
○ Shell (counter- and co-current air-water annular flows with and without surfactants, involving highly turbulent flows, validated against experimental data from the Transient Multiphase Flow Consortium gathered by the Matar Fluids Group in terms of pressure drop and detailed interfacial structure/morphologies);
○ Syngenta (Mixing efficiency of viscoplastic fluids in stirred vessels as a function of material properties, impeller speed and geometry, validated against experimental measurements and visualisations in collaboration

Current Position and Work Experience
Since 2015: Research Associate, Department of Chemical Engineering, Imperial College London. Working under the guidance of Prof. Omar K. Matar. Developed expertise: Massively parallel DNS of fully three-dimensional multiphase flows

2016-2017: Scientific Adviser , Procter & Gamble (P&G) in Brussels (Belgium). Developed expertise: Modelling & simulation of complex industrial applications.

2013-2014: Research Associate at LOMC-CNRS (Waves and Complex Media Laboratory). Univ. of Le Havre, France. Under the supervision of Prof. I. Mutabazi and Dr. J. Peixinho. Project: “Heat transfer by Görtler vortices developed on a wall with finite conductivity".

2013: Research Associate at PMMH and LIMSI-CNRS, under the supervision of Dr. L. Tuckerman, Dr. D. Juric, and Dr. J. Chergui. Project: “Parallel direct numerical simulation of three- dimensional Faraday waves".

2012-2013: Teaching Fellow in mathematics and thermodynamics at Ecole Polytechnique Féminine, and Ecole Centrale d’Electronique, France.

2011-2012: Teaching Fellow in mathematics at the University Paris-Descartes, France.

2010-2011: Teaching Fellow in mathematics at the University Paris-Dauphine, France.

2007-2010: Graduate Teaching Assistant in mathematics at the University Paris-Orsay, France