CTR Seminars Archive 2014

Date: Friday, November 14, 2014
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Elif Karatay, FPCE Postdoctoral Scholar, ME Department
Title: Mass and Momentum Transfer near Interfaces: From bubble surfaces to charged materials

Abstract: The transport phenomena at interfaces often determine the bulk transport rates and thereby determine the performance of both micro- or macro- scale systems. Therefore, a better understanding of the fluid motion and the associated transport processes at these interfaces is essential for further optimization of various micro- and macro- scale technologies. Microfluidics offer an ideal platform allowing for the integration of surfaces with precise and controllable interfaces and direct measurements of transport phenomena driven at these interfaces. Within this context, I will present aspects of my past and ongoing research relevant to interfacial mass and momentum transport. The first part of the talk covers experimental and numerical investigations on the momentum and mass transfer near gas-liquid interfaces established in hydrophobic silicon micro-grooves. The results reveal the impact of bubble geometry on hydrodynamic slippage and mass transfer rates of solutes at curved bubble interfaces. In the second part of the talk, I will present our recent study on chaotic electrokinetic transport near ion selective boundaries with the objective of assessing efficiency of commercial software for prediction of such phenomena. I will present comparisons of detailed statistics against a reference custom-made code that is tailored to the specific physics of electrokinetic transport.  Our results indicate that while accuracy can be guaranteed with proper mesh resolution and avoiding numerical dissipation, commercial solvers are generally at least an order of magnitude slower than custom-made DNS codes. Finally I will conclude with remarks on our ongoing projects aiming at (i) understanding the role of buoyancy forces driven by the variations in salt concentration in electrochemical systems and (ii) direct imaging of electrokinetic chaos in close proximity (~nm) of ion-selective interfaces.

Date: Friday, October 31, 2014
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Yongle Du, Post Doctoral Candidate, Aerospace Engineering, Penn State University
Title: A New CFD/CAA Methodology for Installed Jet Noise Simulations

Abstract: Over a half century of research has yielded a significant reduction of aircraft engine jet noise with innovative noise reduction devices, such as chevrons and fluidic injections. However, experiments have shown that for installed jet engines, their noise reduction effects may be compromised, and the spectral directivity of the radiated noise is altered due to the complex flap-jet interaction, airframe reflection/diffraction, etc. Therefore, understanding these installation effects is crucially important for the noise reduction design at the system level.  
At present, estimate of the installed jet noise relies primarily on semi-empirical models. Although successful for simple isolated jets, the efficient approach using RANS solutions and noise source models showed relatively large errors for installed jets. The more accurate hybrid method combining high-fidelity LES and computational aero-acoustic (CAA) techniques is still very expensive for the complex airframe-jet configurations.

This presentation introduces a new CFD/CAA methodology under development for accurate yet efficient installed jet noise simulations. The CFD/CAA is closely coupled based on the recently developed compact disturbance equations (CDE). The CDE solve the small disturbances about the given base flow. Various reduced governing equations, such as the linearized Navier-Stokes (LNS) and linearized Euler equations (LEE) are included in the same formulation. A seamless switch of the equations can be made between the zones with different flow physics.

An installed jet noise simulation using this methodology is performed in two steps. The steady base flow is obtained first to determine the noise source region and the noise propagation region. Existing methodologies can provide reasonably accurate solutions for complex configurations. A third-party solver and unstructured grids can be used to significantly reduce the computational cost. Second, the coupled CFD/CAA solves the unsteady disturbances in the greatly reduced, simpler sub-domain of interest with an optimal grid design. The simulation restricts the computationally expensive, high-fidelity LES in the confined source region (for example, near the jet shear layer and jet/flap interaction region), and applies the less expensive and more accurate LEE in the vast noise propagation region.
Benchmark tests show that this is an accurate and affordable approach for complex installed jet noise simulations. With good base flow solutions, more accurate boundary treatments and less numerical errors can be achieved in a greatly reduced, simpler CFD/CAA domain in terms of the small disturbances. Furthermore, the reduced equations bring an additional ~35% reduction of the computational cost.

Date: Friday, October 24, 2014
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Manav Vohra, Post-Doctoral Candidate, Mechanical Engineering and Materials Science, Duke University
Title: Development of Reaction Models for Novel Energetic Materials

Abstract: Metallic multilayered systems referred to as energetic materials have attracted immense interest owing to their characteristic reaction properties. Rapid intermixing in the multilayers due to steep concentration gradients and atomic diffusion at length scales of the order of tens of nanometers leads to a large amount of localized heat; thus making such energetic materials suitable for joining applications such as welding and soldering. Moreover, the reaction self-propagates once initiated by means of a high energy source such as an electric spark. Numerous computational studies have focused on capturing the transient reaction phenomena and understanding its dependence on the microstructure and composition of the multilayered systems. During the talk, I plan to discuss my work on developing new models as well as refining the existing reaction models. In particular, the focus would be on calibrating intermixing rates in the multilayers using regression analysis and Bayesian statistical approaches. Recent experimental investigations by Joress et al. [Appl. Phys. Lett. 101.11:111908, 2012] revealed that the oxidation of equimolar Zr-Al multilayers would help extend the duration of heat release by three orders of magnitude as compared to conventional multilayered systems. A simplified computational model to reproduce experimental observations as well as understand the kinetics of oxide layer growth in the equimolar Zr-Al system will also be presented.

Date: Friday, October 17, 2014
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Kwitae Chong, Post-Doctoral Candidate, Mechanical and Aerospace Engineering Department, University of California, Los Angeles
Title: Particle manipulation in viscous streaming

Abstract: A probe of circular cross section, undergoing rectilinear oscillation, creates large-scale steady circulatory cells by viscous streaming. We have shown that inertial particles can be effectively trapped inside these streaming cells, regardless of particle size and density and Reynolds number. we extend this study to various arrangements of oscillating probes. High _delity computations (Viscous Vortex Particle Method) are used to simulate the ow _eld. It is shown that, by controlling the sequence of starting and stopping the oscillation of individual probes, inertial particles can be transported in a predictable manner between trapping points. In order to reduce the considerable expense of generating the ow _eld, we also explore the use of steady Stokes ow to serve as an approximate surrogate for the ow between probes. The boundary conditions for this ow are obtained by matching with the inner Stokes layer solution.

Date: Friday, May 23, 2014
Time: 4:00pm
Location: CTR Conference Room
Speaker: Farid Karimpour, PhD Candidate, Civil and Environmental Engineering Department, Colorado State University
Title: Mixing in stably stratified wall-bounded turbulent flows: Insights and Modeling

Abstract: Stably stratified wall-bounded flows are ubiquitous in nature such as in estuaries, lakes, oceans and atmospheric boundary layer. In such flows, the simultaneous existence of density stratification and solid wall results in anomalous mixing of momentum and active scalar (density) compared to other turbulent flows. Hence, there is no surprise that stratified wall-bounded flows are usually considered as one of the most complex flows. The focus of this study is to understand and model stratified wall-bounded turbulent flows. The equilibrium assumption between the production rate of the turbulent kinetic energy the dissipation rate of the turbulent kinetic energy and the dissipation rate of the turbulent potential energy is invoked to discuss a number of pertinent issues that have direct implications for prediction of the turbulent mixing in stably stratified wall-bounded turbulence. Simple formulations for the flux Richardson number which is commonly considered as a measure of the turbulent mixing in stratified flows and the turbulent viscosity are proposed. Further, RANS simulations of stratified wall-bounded flows are performed. The mixing of density in stratified flows is usually modeled by employing a turbulent Prandtl number which is the linking bridge between the turbulent viscosity and diffusivity. Most parameterizations for Prandtl number are developed based on data obtained from homogeneous shear flows. A one-dimensional stratified turbulent channel flow is modeled and the efficacy of a number of homogeneous turbulent Prandtl number formulations are evaluated. The numerical simulation results highlight the inadequacy of such formulations. We introduce a modified parameterization for Prandtl number that takes into account the inhomogeneity caused by the wall coupled with the effects of density stratification and evaluate its performance. Comparisons with data of direct numerical simulation of stably stratified channel flow show remarkably good agreement.

Date: Friday, May 16, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: High-fidelity numerical simulations of compressible turbulence and mixing generated by hydrodynamic instabilities
Speaker: Pooya Movahed, PhD Candidate, Mechanical Engineering Department, University of Michigan, Ann Arbor

Abstract: The Rayleigh-Taylor (RT) instability occurs in a variety of applications at different scales, ranging from inertial confinement fusion to supernova explosion. In RT unstable configurations, initial perturbations at a heavy-light interface may evolve to a turbulent mixing region, in a process in which the initial potential energy feeds the instability growth. Due to the gravitational field and the density gradient, the resulting turbulence is anisotropic. However, these two effects are generally coupled, such that it is difficult to assess each one individually. The goal of this study is to better understand anisotropy in turbulence through the Rayleigh-Taylor instability using direct numerical simulation (DNS). We use a novel set-up to study the temporal evolution of the mixing region starting from an unperturbed material interface in an existing isotropic field in the presence and absence of gravity. First, we ignore gravity and focus on the temporal evolution of the mixing region due to turbulence diffusion. This set-up allows us to study the role of density gradient across the mixing region individually. At large scales, the mixing region grows self-similarly after an initial transient period; a one-dimensional turbulence-diffusion model in conjunction with Prandtl's mixing length theory is applied to describe the growth of the mixing region. The observed growth exponent tends to 2/7, as expected for Batchelor turbulence based on energy budget arguments for large Reynolds numbers. At small scales, flow isotropy and intermittency are measured. Results suggest that a large density ratio between the two fluids is required to make the velocity field anisotropic at the Taylor microscale, while the flow remains isotropic at the Kolmogorov microscale. Second, we assess the role of gravity in a RT unstable configuration, in comparison to our first set of runs. Now, the baroclinic vorticity due to the gravitational field provides energy driving the initial decaying isotropic field. A comparison of relevant physical quantities regarding isotropy and mixing is made between both cases. The role of different initial most energetic wave numbers of the initial decaying field and Reynolds number are investigated. Current DNS are performed using a high-order accurate minimally dissipative kinetic-energy preserving and interface capturing scheme.

Date: Friday, May 9, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: A fast pressure-correction method for simulating two-fluid flows and DNS of droplet-laden isotropic turbulence
Speaker: Antonino Ferrante, Assistant Professor in Aeronautics and Astronautics, University of Washington

Abstract: Direct numerical simulation (DNS) studies of droplet-laden turbulent flows have mostly been limited to sub-Kolmogorov (d < η) size droplets using the point-particle approach. DNS of finite-size droplets (d > η), characteristic of the size of fuel droplets during secondary atomization, requires fully-resolving the flow inside and around the droplets while accounting for the effects of surface tension. The main goal of the present study is to investigate via DNS the effects of finite-size deformable droplets on decaying isotropic turbulence.

In order to achieve this objective, first, we have developed a three-dimensional volume of fluid (VoF) method for tracking droplets accurately and efficiently in incompressible velocity fields. The novelty of the developed approach is that besides conserving mass globally, a condition not always satisfied by VoF methods, mass conservation is also ensured locally while requiring half the number of advection and reconstruction steps of conventional methods. Then, we have developed and coupled a new pressure-correction method with the VoF method for simulating incompressible two-fluid flows. The method's main advantage is that the variable coefficient Poisson equation that arises in solving the incompressible Navier-Stokes equations for two-fluid flows is reduced to a constant coefficient equation. This equation can then be solved directly using, e.g., the FFT-based parallel Poisson solver that we have developed for petascale supercomputers. For a 1024 mesh, our new pressure-correction method using the FFT-based parallel Poisson solver is ten to forty times faster than the standard pressure-correction method using multigrid. In general, the new pressure-correction method could be coupled with other interface advection methods such as level-set, phase-field, or front-tracking.

Our new pressure-correction/VoF flow solver has been verified up to density and viscosity ratios of 10,000 against theoretical results, validated against experimental results, and shown to conserve mass, momentum, and kinetic energy in the inviscid limit. Finally, I will present results from DNS of non-evaporating droplet-laden isotropic turbulence and the effects of varying the droplet Weber number and the density ratio on the time development of the turbulence kinetic energy budget.

Date: Friday, May 2, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: Linear solver and multiscale modeling methods for enhanced performance in cardiovascular simulations
Speaker: Mahdi Esmaily-Moghadam, PhD Student, Mechanical and Aerospace Engineering, University of California, San Diego
 
Abstract: Numerical simulations in cardiovascular disease present a number of important challenges that necessitate specialized algorithm development for enhanced performance in a high performance computing environment. First, multiscale modeling methods to incorporate dynamic coupling between local hemodynamics and circulatory physiology result in ill-conditioned systems dominated by eigenvalues coming from the coupled boundaries.  Second, computational expense becomes increasingly important when coupling simulations to optimization and uncertainty quantification, and in problems with fluid structure interaction.  Third, high aspect ratio complex geometries present a challenge for conventional linear solvers.   In this talk, I will present our recent developments in numerical algorithms to reduce the cost of solving the linear system of equations arising from incompressible flow and fluid structure interaction problems, and efficiently and implicitly couple flow simulations to reduced order circulatory system models.  First, I will present a novel bi-partitioned algorithm for CFD and FSI problems that is shown to reduce computational cost compared to standard GMRES methods.  Second, I will discuss methods for implicitly coupling flow simulations with reduced order cardiovascular models, and introduce a specialized preconditioner for problems involving coupled boundary conditions.  Third, I will present an efficient data structure for handling iterative solver operations in parallel.  Finally, I will demonstrate the effectiveness of the presented algorithms through several relevant examples.  I will then discuss ongoing research on a new multi-partitioned algorithm and my future plans.

Date: Friday, April 11, 2014
Time: 3:00pm
Location: CTR Conference Room
Title: Thermonuclear Ignition of ICF Capsules: Challenges and opportunities
Speaker: Dr. Baolian Cheng, Los Alamos National Lab
 
Abstract: Ignition is required to make fusion energy a viable alternative energy source. Recently National Ignition Facility (NIF) has achieved record high neutron yield (9e+15) in the high foot inertial confinement fusion (ICF) capsule experiments. However, it is still far away from achieving the gain threshold and ignition. In this talk, I will present the ignition citron, scaling laws, NIF successes, areas to improve and challenges as well as opportunities in both designs and simulations.

Date: Friday, April 4, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: The space-time correlation models for turbulent flows
Speaker: Prof. Guowei He, LNM, Institute of Mechanics, Chinese Academy of Sciences

Abstract: Space-time correlations are simple but fundamental measures of the relationships between turbulent fluctuations in space and time. The space-time correlation models are used to provide the indispensable time scales for two-point closure approach and develop temporally accurate turbulence models. The most well-known model for space-time correlations is Taylor’s frozen flow model. However, it has many limitations such as a weak shear rate and low turbulence intensity. In this talk, I will introduce our recent work on space-time correlations: (1) I will first introduce the EA model for the space-time correlations in turbulent shear flows. The EA model accounts for both propagation and sweeping effects by a second approximation to iso-correleation contours, while Taylor’s model is the first approximation. This model is verified by the DNS of turbulent channel flows and used to transform the temporally varying signals into spatially varying ones in the Rayleigh-Benard experiments. (2) I will further introduce the swept-wave model for compressible turbulence. This model is the extension of the linear-wave propagation model originally developed by Lee, Lele and Moin. It is shown that the temporal decorrelations in dilatational fluctuations are dominated by both random sweeping and wave propagation. The swept-wave model is validated by DNS of compressible and isotropic turbulence and used in the EDQNM approach to derive the scaling of energy spectra.

Date: Friday, March 7, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: Separation of scales in spray combustion
Speaker: Dr. Javier Urzay, Research Associate, Flow Physics and Computational Engineering, Stanford University
 
Abstract: This talk addresses the analytical description of multiscale processes governing spray vaporization and combustion downstream from the injector in liquid-fueled burners. The focus of the presentation is placed on the phenomena involved in the collective combustion of fuel sprays and their aerodynamic interactions with the surrounding flow rather than on the processes occurring at droplet scale. Relevant spray-combustion scales and related dimensionless parameters are presented. Laminar problems are identified that can shed light on modeling different aspects of spray-combustion phenomena. Besides consideration of spherical spray clouds, specific attention is given to group ignition in mixing layers and counterflow spray flames, including inertial effects of droplets with order-unity Stokes numbers. The presentation ends with a brief account of some open problems and modeling challenges in need of additional work.

Date: Friday, February 14, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: Scale Coupling in Richtmyer-Meshkov Flows
Speaker: Prof. Snezhana I. Abarzhi, Carnegie Mellon University
 
Abstract: We systematically study the Richtmyer-Meshkov instability (RMI) induced by strong shocks for fluids with contrasting densities and with small and large amplitude initial perturbations imposed at the fluid interface. The Smoothed particle hydrodynamics code (SPHC) is employed to ensure accurate shock capturing, interface tracking, and accounting for the dissipation processes. Simulations results achieve good agreement with existing experiments and with the theoretical analyses including zero-order theory describing the post-shock background motion of the fluids, linear theory providing RMI growth-rate in a broad range of the Mach and Atwood numbers, weakly nonlinear theory accounting for the effect of the initial perturbation amplitude on RMI growth-rate, and highly nonlinear theory  describing evolution of RM bubble front. We find that for strong-shock-driven RMI the background motion is supersonic, and the interfacial mixing can be sub-sonic or supersonic. Significant part of the shock energy goes
into compression and background motion of the fluids, and only a small portion remains for interfacial mixing. The initial perturbation amplitude appears a key factor of RMI evolution. It strongly influences the dynamics of the interface, in the fluid bulk, and the transmitted shock. In case of large amplitudes, the vector and scalar fields in the fluid bulk are non-uniform. The flow heterogeneities include cumulative reverse jets, checkerboards velocity pattern, shock-focusing effects, and local hot spots with temperature substantially higher than that in the ambient. The dynamics of the nonlinear flow is shown to have a multi-scale character.

Date: Friday, January 10, 2014
Time: 4:00pm
Location: CTR Conference Room
Title: Jet-noise reduction as an inverse problem
Speaker: Dr. Jeonglae Kim, Postdoctoral Research Associate, Mechanical and Aerospace Engineering, Cornell University
 
Abstract: Noise radiation from high temperature jet-engine exhaust is a prime interest to aviation industry and engine manufacturers. However, jet-noise reduction has not been sufficiently guided in general based upon the true physics of turbulent jet; instead, empirical models or extensive parametric studies have been employed. In this study, adjoint-based optimization is utilized so that controls which reduce the sound radiation of a Mach 1.3 turbulent jet are directly explored in conjunction with high-fidelity solutions of Navier–Stokes equations. Jet-noise reduction is formulated as an optimization problem and adjoint of the linearized Navier–Stokes equations is used to provide the control sensitivity. The space–time resolved solutions of high-fidelity simulations are formulated to determine noise-reducing controls, which model heat release effects near the nozzle exit. Controls found by the optimization procedure suppress the jet-noise radiation and make the jet quieter. The simulation generates an unique set of acoustically loud and quiet states of the turbulent jet. Studies demonstrate that radiation of particularly loud acoustic waves is triggered by large-scale interactions of axisymmetric, slowly-propagating disturbances.

Date: Tuesday January 7, 2014
Time: 2:00pm
Location: CTR Conference Room
Title: Environmental fluid mechanics: from density stratified to multi-phase flows
Speaker: Dr. Mona Rahmani, Postdoctoral Student, IFPEN Energies Nouvelles, France
 
Abstract: Large-scale phenomena in geophysical flows provide energy for stirring, mixing, and transport of mass and momentum. The input energy cascades down from macro to meso, and then to micro scales, through chaotic processes. The future challenge for environmental fluid dynamists is to understand the interaction of these multi-scale processes. Environmental flows commonly exhibit these features: density stratification caused by variations in temperature, salinity, or concentration of suspended particles, and addition of a secondary phase(s) such as solid particles, vegetation, liquid droplets, or air bubbles to the primary phase of liquid or gas. This talk addresses some topics in environmental fluid mechanics of density stratified and multi-phase flows, that involve studying natural processes at different scales. The first part of the talk focuses on direct numerical simulations of mixing caused by shear instabilities in density stratified flows. As the Reynolds number increases a transition in the overall amount of mixing is found, which is in agreement with previous experimental studies. The effect of Prandtl number on mixing is studied to understand the characteristics of high Prandtl number mixing events in the ocean; these cases have usually been approximated by low Prandtl number simulations. The increase in the Prandtl number has some significant implications for the evolution of the flow, the time variation of mixing properties, and the overall mixing. The second part of the talk focuses on fully resolved simulations of particulate flows. Path instabilities of the free motion of spherical and angular particles are examined for increasing Reynolds numbers. The results reveal mechanisms of path instabilities for angular particles that are different from those for spherical ones. Some results of ongoing research on suspension of spherical particles is also discussed. The goal is to understand how the micro-scale interactions of the particles influences the properties of the suspension as a whole.