Gustaaf (Guus) Jacobs

Research

Summary

Engineering: Combustors, Drag Reduction (Automotive/Aeronautical), Renewable Energy (Wind Turbines), High Power Microwaves.
Physics: Multi-Phase Flows, Multi-Scale Physics, Flow Control, Unsteady Separation, Compressible Turbulence, Combustion, Electromagnetics, Plasmas.
Applied Mathematics: High-order Multidomain, Discontinous Galerkin, Spectral h/p Methods, Particle-In-Cell Methods, High Performance and Parallel Computing.

Projects:

Multi-scale Methods for Computation of Particle-Laden Flows with Shocks

Multi-scale computational methods are developed for the computation of particle- and droplet-laden flow with shocks. The methods increase efficiency and fideltiy over established macro-scale methods. Through surrogate models, meso-scale codes are coupled with macro-scale models. Uncertainty is quantified in the coupling between scales. The figure and animation (click on the figure) shows results of a macro-scale computation using a WENO based particle-mesh method of a shock running into a particle cloud.
Shock-Particle interaction
The figure shows the particle dispersion at late times superimposed on the vorticity contours. In the animation particles are superimposed on the density contours. The high-order method was shown to be extremely efficient in capturing the small scale features in th particle-laden flow, while sharply capturing the shocks.
Collaborators: ` Dr. H.S. Udaykumar (University of Iowa), Dr. W.S. Don (Brown University),

Droplet-Laden Flow Separation Analysis in the Lagrangian Framework

Droplet-laden flow separation is analyzed in the Lagrangian framework, i.e. the frame moving with fluid particles and droplets, as opposed to the more common analysis of streamlines and wall friction fields in the Eulerian frame. In unsteady and three-dimensional flows, fluid particle separation is suprisingly different from the Eulerian streamline separation. In a quasi-periodic unsteady flow, fluid particles eject away from a fixed wall location and form an oscillating material line into the flow that is seemingly unrelated to the unsteady streamline pattern. This 'fixed separation' has been validated in a rotor-oscillator experiment as shown in the animation. The fluid particle separation location in 2D and 3D is, however, simply identified with time history of wall data including the wall shear stress and wall pressure data. In flows with large turbulent fluctuations the fluid particles separate nearly parallel to the wall and eject away from the wall into the flow away fromt the fixed separation location. The separation material line is then better described as moving. We have recently discovered that finite-sized droplets may still separate in a fixed manner in these moving carrier flows. For more details see the recent news articles:
Tecplot: plot of the month
MIT solves 100-year-old engineering problem
and webinar given at Tecplot ( link ).
Collaborators: Dr. G. Haller (MIT), A. Surana (MIT)

Droplet-Laden, Reacting, Turbulent Flow in Spray Combustors

The long term objective of this project is the elucidation of the combustion physics in liquid-fuel combustors using simulation tools. To this end, we compute the droplet-laden, reacting flow in combustor geometries using Direct Numerical Simulation (DNS) and Large Eddy Simulations (LES) to solve the gas dynamics coupled with particle methods to trace the droplet dynamics. Large-scale computations have been performed of the droplet-laden flow in a combustor geometry with a sudden expansion as shown in the figure. Compressibility effects have been identified to reduce the recirculation zone behind the step. Control of the cold flow using a suction slot at the step corner has been shown to be an effective destabilizer of the shear layer and to enhance mixing behind the step. Droplet mixing and deposition is strongly dependent on the injection location in the boundary layer. Current efforts focus on further parametric studies to identify the combustion physics, and computations of more complex combustor geometries.
Collaborators: Dr. F. Mashayek (University of Illinois at Chicago), K. Sengupta (University of Illinois at Chicago)