CTR Seminars Archive 2007
Date: December 14, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Riccardo Rossi, Visiting Scholar, Universita' degli Studi di Bologna, Italy
Title: Numerical simulation of scalar dispersion in complex flows
Abstract: The analysis of scalar dispersion in turbulent flows is relevant to a broad range of applications. In the last fifty years experimental and numerical investigations of canonical flows have dealt with the improvement of the theoretical background and subsequently with the development of reliable low-order models. However, although these techniques have been successfully employed in basic flow configurations, it is generally accepted that LES should provide a significant improvement in the framework of complex flows. The first part of the talk will give a short overview of scalar transport phenomena and the numerical challenges associated with high-fidelity numerical simulations of turbulent scalar transport by presenting results from DNS of turbulent channel flows. In spite of the linear character of the scalar transport equation, the results show that the scalar problem is far from being a footnote to turbulence research. In the second part of the talk the numerical simulation of the turbulent dispersion of a passive scalar from a line source downstream of a surface-mounted obstacle will be discussed. A RANS-based analysis has been initially carried out using the standard k-epsilon model, the standard k-omega model and a RST closure. A direct numerical simulation is then performed to investigate the dynamics of the scalar wake and, furthermore, to provide reliable results for the Reynolds stress and turbulent scalar fluxes. The predicted profiles of scalar statistics at two streamwise locations are compared to experimental measurements available in the literature. The most interesting result obtained from the RANS-based analysis is that a very accurate prediction of the local mean velocity profiles does not represent the key requirement in the present computations. The turbulent scalar diffusivity has been found the most important parameter indeed. The very detailed insight given by the direct numerical simulations leads finally to the conclusion that the entrainment of the scalar plume above the line source and the subsequent development of the scalar wake are primarily dominated by the breakdown of the shear layer generated at the upper leading edge of the obstacle.
Date: December 7, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Rene Pecnik, CTR Postdoctoral Scholar
Title: Bypass transition modeling and simulations of compressible unsteady turbomachinery flows with Reynolds-averaged Navier-Stokes approach
Abstract: This seminar will be split into two parts, the first one deals with bypass transition modeling, and the second part shows unsteady results of transonic turbomachinery flows. Nearly all kinds of fluid motion in our surroundings are turbulent by nature which makes turbulence modeling of major interest for engineering applications. While popular RANS turbulence models are able to represent most of the important turbulence phenomena correctly, they suffer from the inability to predict laminar-to-turbulent boundary layer transition. The accurate numerical simulation of turbomachinery flows often depends on the reliable prediction of transition. The first part of this talk deals with modeling bypass transition in thermal turbomachines under the influence of high freestream turbulence intensities, by means of different approaches: firstly empirical transition models based on an intermittency factor and secondly turbulence models having the ability to predict transition. The results show that empirical models are able to predict transition reasonably. For a broader application range, models based on a solid physical understanding of transition have to be developed. The herein investigated turbulence closures, namely the V2F and the turbulent potential model, are directly related to full Reynolds stress transport models and keep the essential information of turbulent anisotropy by means of the pressure strain correlation. This seems to be the key to predict bypass transition. A further topic of interest in turbomachinery research and design of future aero-engines is unsteady fluid phenomena in high pressure turbines where the main source of large scale unsteadiness is due to the interaction between stationary and rotating blade rows. Further the high stage loading and the low aspect ratio typical of high pressure bladings induce transonic flows with three-dimensional shock surfaces traveling through the flow passages and strong secondary vortices. To provide a detailed insight of these unsteady phenomena the second part of the talk will show unsteady three-dimensional computations of the transonic flow through such turbines. All simulations presented in this talk have been performed with an own developed implicit compressible flow code.
Date: November 30, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Thomas Lederlin, CTR Postdoctoral Scholar
Title: Design and experimental/numerical study of a control system for gaseous-jet trajectory and mixing in oxygen-fueled burners
Abstract: This work deals with the control of multi-jet flames for an application to oxy-fuel multi-jet burners. The generic type of control that is studied here is an axisymmetric jet being vectored by a transverse fluid injection. An actuator, designed at Institut de Mécanique des Fluides de Toulouse (IMFT), carries out this injection and gives a large amplitude to the jet deviation. The aerodynamics of the flow issuing from a double-jet injector, geared with actuators, is then studied experimentally and numerically. Two different flow configurations prove the ability of the actuator to enhance mixing and to control the trajectory of the global flow. The experimental visualization and quantitative measurements, achieved with hot-wire velocimetry, are retrieved well by large-eddy simulations. Reactive calculations on a three-jet configuration are also undertaken. When compared to experimental results, these calculations show the effect of jet incidence on the flame structure and prove the ability of LES to capture such phenomena. At the industrial level, a 1-MW pilot burner has been outfitted with the IMFT actuators. Tests of this burner prove the effectiveness of the actuators to orient the heat transfer from the flame to a load, and exhibit better overall performance than a "straight-flame" burner.
Date: November 9, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Professor Bahram Moshfegh, Linkoping University, Sweden
Title: LES predictions of an impinging jet in a cross-flow on a heated wall-mounted cube
Abstract: The current trends towards the miniaturization of electronic devices and increasing processing speed are resulting in a steady increase in the amount of heat dissipated per unit surface area or unit volume of the electronic components. The performance of electronic systems to a great extent is limited by the possibility of maintaining the electronic systems at sufficiently low temperatures. This fact makes the thermal management of electronics a great challenge for the thermal engineers. Forced channel flow is frequently used to remove heat at the walls of the channel where Printed Circuit Boards (PCB) are located. The PCB will contain one or a very few high heat dissipating components and many more peripheral lower dissipating electronic components. Thus the overall cooling strategy must not only match the overall power dissipation load, but also address the requirements of the "hot" components. By combating the whole thermal load with forced channel flow, excessive flow rates will be required. In this study we will explore the combined use of forced convection channel flow between printed circuit boards to provide the overall thermal management and with "spot" cooling using impingement jet on hot components. The objective of this study is to perform a Large Eddy Simulation (LES) in order to predict the mean velocity field, the turbulence characteristics and the heat transfer rate of the impinging jet in cross-flow configuration on a heated wall-mounted cube. The results from the LES are compared with a Reynolds stress model (RSM) and a v2-f model as well as against earlier measurements with identical set-up. The results also indicate that the cooling performance on the top of the component can be increased by over 300%.
Date: November 2, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Kun Luo, CTR Postdoctoral Scholar
Title: Direct numerical simulation of particle dispersion and particle-fluid interactions in shear flows
Abstract: Although the preferential concentration for intermediate particles has been well-known for a long time, no state-of-the-art model has been developed to describe and predict it. As the first step towards this direction, a DNS study of particle dispersion in a transitional jet is initially introduced. The effects on particle dispersion by turbulence transition and the correlations between particle distribution and fluid pressure are investigated. It is found that for particles at the intermediate Stokes number, a transitional dispersion from non-uniform (Preferential concentration) to uniform occurs. This effect should be taken into account when modelling multi-phase reacting flows and combustion. The distribution of particles in the flow field also correlates well with the distribution of the Laplacian of pressure. But from the statistical point of view, the fraction of particle number distributed within different flow zones characterized by the Laplacian of pressure becomes time-independent and size-independent when the flow is developed enough. In the second part, a modified immersed boundary method and its applications will be talked. Based on boundary theory, a non-linear interpolating scheme was proposed to get the fluid velocity near the interfacial boundary. This scheme contains some physics and can get favorable results compared to experimental data. Some potential incorporation with modelling multi-phase reacting flows and combustion will be discussed too.
Date: October 19, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Olaf Marxen, CTR Postdoctoral Scholar
Title: Towards numerical simulation of hypersonic boundary-layer instability including high-temperature gas effects
Abstract: One of the most critical components of a planetary exploration vehicle is the thermal protection system designed to withstand extremely large heat loads. Turbulent flows induce a much higher heat load in comparison to laminar flows. The occurrence and location of laminar-turbulent transition is therefore a key factor in heat-shield design. Fundamental physical processes leading to transition on smooth surfaces are not well understood, but even less so for localized roughness elements at the wall, such as bolts or cavities present on typical heat shields. In comparison to low speed transitional boundary layers, hypersonic flows can show conceptually different phenomena. Among these are shocklets and multiple instability modes. These additional complexities can make fundamental investigations more demanding and engineering predictions less accurate. Within the present work, time-accurate numerical solutions to the compressible Navier-Stokes equations are obtained by a high-order finite-difference code. The fluid is modelled as calorically perfect gas (constant specific heat), thermally perfect gas (variable specific heat), or chemically reacting gas in equilibrium (infinitely fast reaction rate). The basic configuration is a flat plate boundary layer with an adiabatic wall. Two different Mach numbers, Ma=4.8 and Ma=10, are considered. For the Ma=4.8 case, we study the influence of a localized roughness element. This element takes the form of a small, two-dimensional hump. For the Ma=10 case, we examine high-temperature gas effects. Small perturbations at a fixed frequency are triggered at the wall via blowing and suction. Due to the instability of the boundary layer, these perturbations grow while convecting downstream. If the amplification is sufficiently strong and occurs for a long downstream distance, this can lead to breakdown to turbulence.
Date: October 3, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Prof. C. Wagner, Institute for Aerodynamics and Flow Technology, German Aerospace Center, Gottingen, Germany
Title: Analysis of coherent structures and the heat transport in turbulent Rayleigh-Benard convection based on Direct Numerical Simulations and Large-Eddy Simulations
Abstract: Direct Numerical Simulations (DNS) and Large-Eddy Simulations (LES) of turbulent Rayleigh-Benard convection in cylindrical and cuboidal containers filled with air have been performed in a wide Rayleigh number regime. The LES were conducted with the tensor diffusivity model or a dynamic scale similarity model. The performance of the latter will be demonstrated presenting results obtained in a LES of a turbulent channel flow. Further, a new method to detect and analyse coherent structures in turbulent Rayleigh-Benard convection which is based on a thermal dissipation rate analysis will be presented and discussed. Analysing the generated DNS and LES data allowed to identify coherent structures and to determine their role on the turbulent vertical heat flux for different Rayleigh numbers. Finally, global and local heat flux distributions will be discussed in the investigated Rayleigh number regime.
Date: September 7, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Rob J.M. Bastiaans, Eindhoven University of Technology, The Netherlands
Title: DNS, LES and analysis of premixed methane and hydrogen combustion with flamelet generated manifolds.
Abstract: Hydrogen is considered more and more as a fuel for future combustion systems, especially for power production. Compared to methane there are less elementary reactions but the complexity is higher due to large differences in the transport of mass and energy (Lewis number) and preferential diffusion effects. Thermo-diffusive instabilities and turbulence induced stretch effects can change the turbulent fuel consumption considerably, even in the subgrid scales. It is our goal to extend the present flamelet generated manifolds method to more complex fuels and combustion concepts. The method was validated for several cases with methane as a fuel. For this case (Le=1) a strong stretch theory was rigorously derived and numerically validated as well. We want to extend this to hydrogen cases and mixtures of methane with hydrogen. To that end DNS's are needed in which the basic ingredients of the physics of hydrogen combustion are included. The present state of the art is either performing 3D DNS's with single step chemistry or doing 2D DNS with detailed chemistry. A flamelet method is developed for this purpose. Currently it is being analysed with respect to its thermo-diffusive, preferential diffusion and stretch behavior.
Date: August 24, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Sathyanarayana Ayyalasomayajula, Sibley School of Mechanical and Aerospace Engineering, Cornell University
Title: Eulerian and Lagrangian investigations of simple turbulent flows
Abstract: Investigations of turbulent flow take two approaches: measurements at a fixed point in space (Eulerian) and measurements following a fluid parcel (Lagrangian). In this talk we present results of experimental measurements and models of simple turbulent flows. We investigate several aspects of high Reynolds number turbulence using both approaches. We first present an Eulerian study of high-Reynolds number, homogeneous, isotropic turbulence that has been strained via an axi-symmetric contraction. We study the effect of strain on the turbulence using detailed hot-wire measurements, and we compare the results with Rapid Distortion Theory (RDT). The effect of strain on intermittency as well as the structure of turbulence is also presented. Second, we describe experimental measurements of inertial particle Lagrangian accelerations in high-Reynolds number, homogeneous, isotropic turbulence in wind tunnels. These are the first measurements of inertial particle accelerations in a well-documented high-Reynolds number wind-tunnel flow [Ayyalasomayajula et al Phys. Rev. Lett 97, 144507 (2006)]. Using a high-speed camera that moves with the mean speed of the flow, inertial particle Lagrangian trajectories are captured and their accelerations are calculated. The acceleration probability distribution functions (PDFs) are compared with those of recent DNS where we find good agreement. We propose a new model called the Vortex model, which is used to study the inertial particles in turbulent-like flows. We use this model to study several mechanisms through which attenuation of inertial particle accelerations & acceleration PDFs may occur in real turbulent flows. We also compare the results of the Vortex model with those of the DNS simulations of Bec et al. [J. Fluid Mech 550 (2006)]. Finally, we discuss the range of applicability and shortcomings of stochastic acceleration models for inertial particle modeling.
Date: August 10, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Luca Brandt, Linne' Flow Centre, Royal Institute of Technology, Stockholm, Sweden
Title: Transition delay in boundary layer flows: Stabilization of Tollmien-Schlichting waves by finite amplitude streaks
Abstract: The reduction and control of the viscous drag force exerted on thin bodies moving in a fluid is of great technical interest. Several active and passive methods to achieve a delay of laminar-turbulent thin bodies moving in a fluid is of great technical interest. Several active and passive methods to achieve a delay of laminar-turbulent transition have been developed in the past. The study by Cossu and Brandt (Phys. Fluids (14), L57, 2002) showed the stabilization of the Tollmien-Schlichting (TS) waves by steady streaks of finite amplitude in the Blasius boundary layer. In the presence of streaks, the unstable TS-waves evolve from two-dimensional waves to spanwise modulated waves, referred to as streaky TS-waves. They have similar phase speed as their two-dimensional counterpart and are less unstable. The experiments by Fransson et al. (Phys. Fluids 17-054110, 2004) confirmed the theoretical predictions and demonstrated that such a stabilizing effect can indeed lead to transition delay (Phys. Rev. Lett. 96-064501, 2006). After revisiting these previous experiments, the talk will focus on a recent numerical study of such stabilization in a realistic framework where the effect of streaks of varying amplitude, spacing, and the corresponding sensitivity of the transition delay is investigated. In particular, the evolution of perturbations at low streak amplitudes is examined. To successfully reproduce and extend the experimental findings the use of large-eddy simulation (LES) is necessary to obtain accurate results at the available computational power. For this purpose the ADM-RT model is employed, which was found to be well suited for spectral simulations of transitional flows (Schlatter et al. Int. J. Heat Fluid Flow, 25(3):549, 2004).
Date: June 22, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Jan Nordstrom, Ph.D., Visiting Professor at CTR
Title: The use of modelproblems in CFD
Abstract: The equations governing fluid flow are very complex and most models are non-linear. To gain insight and get ideas on how to analyse/improve solution methods for these equations, simplified model problems are a big help. In this talk we will discuss these matters and show some examples on successful use of model problems.
Date: June 15, 2007
Time: 9:15 AM
Location: CTR Conference Room
Speaker: Andrew Siegel, Ph.D., University of Chicago & Argonne National Lab
Title: Fast Reactor Simulation (please note the time)
Abstract: An informal talk introducing the key physics of fast reactor design and the simulation state-of-the-art and challenges.
Date: June 8, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Guillaume Balarac, Ph.D., CTR Postdoctoral Fellow
Title: Numerical study of vortex dynamics and mixing in turbulent coaxial jets
Abstract: This work is devoted to the study of round coaxial jets with high velocity ratios using numerical simulations. First, direct numerical simulations (DNS) of coaxial jets with moderate Reynolds numbers are carried out to study the transition towards fully-developed turbulence state. The development of the inner and outer Kelvin-Helmholtz rings, arising from the initial instabilities is not independent. The inner rings are controlled by the outer shear layer as they travel downstream. Moreover, it was observed that coaxial jets develop a back-flow region when the ratio between the outer and the inner jet velocities exceeds a critical value. The appearance conditions and the influence of the back-flow region have been studied. The mixing properties of these jets are then studied in seeding a passive tracer in the outer annular jet. The coherent structures control the mixing process. The streamwise vortices improve therefore the mixing by ejections of tracer around the jet. Thus, the flow configurations favoring the intense or early streamwise vortices allow to improve the mixing. Large-eddy simulations (LES) are also performed to understand the modifications on the vortex dynamic and the mixing process due to the Reynolds number value. Finally, following the conclusions of previous points, an active control process based on azimuthal forcing of the outer shear layer is used to improve the mixing efficiency.
Date: May 25, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Liangyu Wang, Ph.D., CTR Postdoctoral Fellow
Title: Thermal Radiation Modeling in Turbulent Combustion Simulations
Abstract: In fires and combustion simulations, especially in computational fluid dynamics (CFD) based simulations, accurate and efficient modeling of thermal radiation has been an issue. The accuracy of predicting pollutant emissions, such as NOx and soot, depends on the accuracy of radiation modeling, which strongly affects flame temperature distributions. There are four major difficulties in radiation modeling: the determination of radiative properties, the solution of radiative transfer equation (RTE), the spectral integration, and in the case of turbulent combustion, the treatment of turbulence/radiation interactions. Strategies to overcome these difficulties have been developed and will be discussed. One of the key strategies is the Full-Spectrum k-distribution (FSK) method. The implementation of the FSK method greatly facilitates the nongray radiation modeling and the employment of accurate Photon Monte Carlo (PMC) method in CFD simulations. The further integration of the PMC method with the transported probability density function (TPDF) method provides an exact treatment of turbulence/radiation interactions without any approximation. The development of the FSK method for inhomogeneous media will be briefly described. This presentation will show the state-of-the-art radiation modeling in turbulent combustion simulations.
Date: May 11, 2007
Time: 11.00am (Please note the time)
Location: CTR Conference Room
Speaker: Guy Dimonte, Ph.D., Los Alamos National Laboratory
Title: K-L turbulence model and validation experiments for the Rayleigh-Taylor and Richtmyer-Meshkov instabilities
Abstract: A turbulence model is developed to described the self-similar growth of the Rayleigh-Taylor (RT) and Richtmyer-Meshkov (RM)instabilities. The model describes the dominant eddies in the mixing zone with evolutionary equations for their characteristic dimension L and energy per unit mass K a V^2/2. The equations are based on the successful buoyancy-drag models for RT and RM flows, but constructed only with local parameters so that it can be applied to multi-dimensional flows with multiple shells of materials. The model has several unknown coefficients that are determined by comparing analytical and numerical solutions with RT and RM experiments. Future validation experiments will also be discussed involving RT unstable flows in cylindrical geometry and with combustion. This experiment is intended to be a chemical analogue to the thermonuclear ignition experiment on the National Ignition Facility at LLNL in order to study turbulent mixing with 3x energy gain.
Date: May 3, 2007 (Thursday)
Time: 4:00pm
Location: CTR Conference Room
Speaker: Xiaowen Wang, Ph.D., Graduate Research Assistant at University of California, Los Angeles
Title: Numerical Simulations of Supersonic Boundary-Layer Stability and Receptivity
Abstract: The stability and receptivity of supersonic boundary-layer flows to various wall disturbances, studied by numerical simulations, are reported in this presentation. The characteristics of boundary-layer wave modes are identified and evaluated by comparing the results of linear stability theory (LST) and numerical simulations. For the case of a Mach 8 flow over a sharp wedge to two-dimensional wall blowing-suction, the most important phenomenon consistently shown by series of numerical simulations is that the synchronization point of mode F and mode S plays an important role in the excitation of mode S by wall blowing-suction, i.e., mode S is strongly excited only when the blowing-suction actuator is located upstream of the synchronization point. When the forcing actuator is downstream of the synchronization point, there is very little excitation of mode S, despite the fact that the blowing-suction actuator is still located within the unstable region of mode S. For another case of a Mach 5.92 flow over a flat plate to two-dimensional wall perturbations including oscillation, blowing-suction, temperature perturbation, and to three-dimensional stationary roughness elements, the numerical results show that all two-dimensional wall perturbations eventually result in the same type of instability wave (mode S) in the boundary layer. The hypersonic boundary-layer flow is most sensitive to wall blowing-suction and least sensitive to wall temperature perturbation. Counter rotating streamwise vortices are excited by three-dimensional stationary roughness element.
Date: April 27, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Madhusudan Gurpura Pai, Ph.D., Mechanical Engineering at Iowa State University
Title: Multiscale Probability Density Function Modeling of Multiphase Flows
Abstract: Many engineering applications such as fuel sprays and fluidized beds involve multiphase turbulent reactive flows. Multiphase flows lend themselves naturally to a statistical description. The nonlinear dependence of the unclosed terms on known quantities in the governing equations for a multiphase flow motivates the need for a probability density function (pdf) closure in preference to moment closures. Multiscale models for the unclosed terms are in turn necessary to capture the complex interaction of the dispersed particles with the carrier-phase turbulence. Predicted trends of phasic turbulent kinetic energy (TKE) from a widely-used Lagrangian-Eulerian drag model are compared with DNS results of a canonical two-phase flow for varying Stokes number. This omparison reveals the inadequacy of the particle response timescale to capture the complex multiscale particle-turbulence interaction. When the particle response timescale is replaced by a multiscale interaction (MI) timescale, the predicted trends match the DNS results, thus validating the need for a multiscale model. Howurbulent reactive flows. Multiphase flows lend themselves naturally to a statistical description. The nonlinear dependence of the unclosed terms on known quantities in the governing equations for a multiphase flow motivates the need for a probability density function (pdf) closure in preference to moment closures. Multiscale models for the unclosed terms are in turn necessary to capture the complex interaction of the dispersed particles with the carrier-phase turbulence. Predicted trends of phasic turbulent kinetic energy (TKE) from a widely-used Lagrangian-Eulerian drag model are compared with DNS results of a canonical two-phase flow for varying Stokes number. This omparison reveals the inadequacy of the particle response timescale to capture the complex multiscale particle-turbulence interaction. When the particle response timescale is replaced by a multiscale interaction (MI) timescale, the predicted trends match the DNS results, thus validating the need for a multiscale model. However, in a real two-phase flow, particle dispersion is coupled to the interphase TKE transfer. DNS of canonical particle-laden flows report that phasic TKE and particle dispersion evolve on timescales that behave differently with Stokes number. A new Lagrangian-Lagrangian model called the dual-timescale Langevin model, which includes the MI timescale, is proposed. This model can simultaneously capture the disparate Stokes number trends in the evolution of TKE and particle dispersion statistics in both phases. Model predictions of two-phase flow statistics are shown to match well with DNS results. In order to extend the above advances in modeling to inhomogeneous flows, a concomitant improvement in our understanding of the underlying theoretical framework for multiphase flows is essential. A framework for the consistent statistical representation of two-phase flows based on a transported pdf formalism is presented. The availability of a consistent statistical representation of two-phase flows motivates the development of a new class of combined EE-LE implementations that have promising applications in a variety of multiphase turbulent reactive flows.
Date: April 13, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Panagiotis Stinis, Ph.D, Postdoctoral Fellow at Lawrence Berkeley National Laboratory and UC Berkeley
Title: Long Memory Mori Zwanzig Models for the Euler Equations
Abstract: A long memory model for dimensional reduction, known as the t-model, is derived through the Mori-Zwanzig formalism of irreversible statistical mechanics. The model is applied to the estimation of the rate of decay of solutions of the Burgers equation and of the Euler equations in two and three space dimensions. In the Burgers case, the model captures the rate of decay exactly. For the Euler equations in two space dimensions, the model preserves energy as it should. In three dimensions, we find a power law decay in time and observe a temporal intermittency. If time permits, we will discuss briefly a hierarchy of Mori-Zwanzig models in which the t-model is the first one.
Date: April 6, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Laetitia Jacoutot, CTR Postdoctoral Fellow
Title: Numerical Modeling of Coupled Phenomena within Molten Glass Bath Heated by Induction
Abstract: This presentation reports on a new vitrification process developed by the French Atomic Energy Commission (CEA, Marcoule). This process is used for the treatment of high activity nuclear waste. It consists in incorporating the ultimate waste in molten glass. This process is characterized by the cooling of all the metal walls and by currents directly induced inside the molten glass. In addition, a mechanical stirring device is used to homogenize the molten glass. The goal of this study was to develop numerical tools to understand phenomena which take place within the bath and which involve thermal, hydrodynamic and electromagnetic aspects. The main difficulties encountered are related to two complications: 1) the large thermal variations of the physical properties of the glass, requiring coupling of the electromagnetic, hydrodynamic and thermal aspects and 2) the asymmetry created by the stirring systems used to homogenize the molten glass bath. Three geometries have been calculated in order to overcome the difficulties one by one: 2d-axisymmetric (without stirrer), quasi 2d-axisymmetric (with vertical stirrer) and 3D configuration (with actual non vertical stirrer). All the calculations have been validated using experimental results obtained from pilot vitrification facilities.
Date: March 16, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Takao Suzuki, Associate Professor, Senior Visiting Scholar at Center for Turbulence Research
Title: Hybrid Unsteady Flow Simulation Combining DNS and PTV
Abstract: In this seminar, I introduce hybrid unsteady flow simulation combining direct numerical simulation (DNS) and particle tracking velocimetry (PTV). We rectify instantaneous PTV velocity fields in a least square sense so that they satisfy the equation of continuity, and feed them to the DNS by equating the computational time step with the frame rate of the PTV measurement system. As a result, we can reconstruct unsteady velocity fields that satisfy the governing equations based on experimental data, with the resolution comparable to numerical simulation. To demonstrate the capabilities of the hybrid algorithm, we investigate a nominally two-dimensional flow past the NACA0012 airfoil at a low Reynolds number (Re=1300). Particle velocities are acquired on a laser sheet in a water tunnel, and an unsteady flow for a deep stall condition (alpha = 15 deg.) is reconstructed with the hybrid algorithm solving incompressible DNS in two dimensions. Intensive alternate vortex shedding, which is predicted by the two-dimensional DNS, is substantially suppressed in the hybrid simulation, and the resultant flow field is similar to the PTV velocity field, which is governed by three-dimensional fluid dynamics. Using the hybrid algorithm, unsteady pressure distribution can also be solved, and the lift and drag coefficients are calculated based on the PTV measurement. By performing hybrid simulations at higher Reynolds numbers, we evaluate the accuracy of this technique compared with an experimental result. In addition, we attempt to identify the motion that originates three-dimensional flow patterns by highlighting the deviation of the PTV velocity field from the two-dimensional governing equations at each snapshot. The results demonstrate that the intensive spots of the deviation appear in the regions in which three-dimensional instabilities are induced in the shear layer.
Date: March 2, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Soshi Kawai, CTR Postdoctoral Fellow
Title: Compressible Large-Eddy Simulation for Aerospace Applications: Transition and Aeroacoustics
Abstract: Investigations of compressible large-eddy simulation (LES) for aerospace applications, subsonic transitional boundary layer and supersonic overexpanded jet aeroacoustics, are introduced in this presentation. Numerical schemes used for compressible LES are mainly divided by two parts whether the flow involves shock waves or not. In the first part, compressible LES is applied to the subsonic transitional boundary layer without shock waves using compact differencing and spatial filtering schemes. Some guidelines regarding how the LES can be used for the analysis of the transitional boundary layers are shown. Particular attentions are paid to evaluating the sensitivities of simulation parameters, such as grid resolution and the filtering parameter, to the transitional flowfields associating with key flow physics of the transition. In the second part, high-order weighted compact nonlinear scheme (WCNS) is extended to the simulation on curvilinear grids and applied to the noise simulation from rocket engine plume involving shock waves. Uniform preserving and vortex preserving properties of WCNS on wavy grids are evaluated by comparing with weighted essential no oscillatory (WENO) scheme. Then, aeroacoustic mechanisms of an overexpanded supersonic jet impinging on a flat plate with hole are qualitatively discussed from axisymmetric simulations using WCNS.
Date: February 16, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Anne Birbaud, CTR Postdoctoral Fellow
Title: Dynamics of Interactions Driving Combustion Instabilities
Abstract: This presentation concerns the fundamental mechanisms responsible for the resonant coupling between flow perturbations, combustion dynamics and acoustics. It focuses on premixed flame response to flow non-uniformities. The objective is to investigate and analyze these interactions to achieve a better understanding of combustion instabilities. Two approaches are proposed to characterize flame dynamics in various situations. The first consists in an experimental investigation of simple but well controlled burner geometries in order to analyze the parameters governing flame response to the upstream acoustic excitation. The second approach uses numerical simulations of some experimental configurations to obtain complementary data on flame dynamics. These combined tools permitted to characterize the effect of various parameters such as flame geometry, excitation frequency and amplitude, lateral confinement and equivalence ratio fluctuations on the flame transfer function. Results indicate that the flame response is strongly non-linear and is very sensitive to flow perturbations and flame geometry. The talk is divided into three parts : first, upstream flow dynamics of a free jet and a conical premixed flame submitted to upstream acoustic excitation is discussed. In the second part, the effect of lateral confinement ratio is investigated on inverted conical premixed flame response to flow perturbations. Results serve to analyze the system stability using the compact acoustic elements method. The last part concerns flame response to equivalence ratio fluctuations. Numerical simulations indicate that laminar flame speed variations induce flame front oscillations, which in turn influence the velocity field in the fresh gases. For sufficiently high modulation amplitudes, the flame responds at the sub harmonic of the excitation frequency.
Date: February 2, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Alireza Doostan, CTR Postdoctoral Fellow
Title: Probabilistic Construction and Numerical Analysis of Model Verification and Validation
Abstract: In this presentation, some recent developments in verification and validation of predictive stochastic models are introduced. Verification is a mathematical concept which aims at assessing the accuracy of the solution of a given computational simulation compare to sufficiently accurate or analytical solutions. Validation, on the other hand, is a physics-based issue that aims at appraising the accuracy of a computational simulation compare to experimental data. The proposed developments are based on an approximation-theoretic representation of random quantities defining computational models both at system and response levels. In particular, three types of problems will be addressed. First, a-priori and a-posteriori error analysis of spectral stochastic Galerkin schemes are briefly discussed. Second, a statistical procedure will be developed in order to calibrate the uncertainty associated with parameters of a predictive model from experimental or model-based measurements. An important feature of such data-driven characterization algorithm, is in its ability to simultaneously represent both the intrinsic uncertainty and also the uncertainty due to data limitation. Third, a stochastic model reduction technique is proposed in order to increase the computational efficiency of spectral stochastic Galerkin schemes for the solution of complex stochastic systems. In particular, an algorithm is developed for the efficient characterization of a lower dimensional manifold occupied by the solution to a stochastic partial differential equation in the Fock space associated with the Wiener chaos. While the second part of the talk is essential in model validation phase, the first part is particularly important as it provides one with basic components of the verification phase.
Date: January 19, 2007
Time: 4:00pm
Location: CTR Conference Room
Speaker: Hossam A. El-Asrag, Ph.D.
Title: Large Eddy Simulation Subgrid Model for Soot Prediction
Abstract: A new soot formation subgrid model is developed and reported here. The new model is designed to be used within the context of the Large Eddy Simulation (LES) framework, combined with Linear Eddy Mixing (LEM) as a subgrid combustion model. The final model can be applied equally to premixed and non-premixed fames over any required geometry and flow conditions in the free, the transition, and the continuum regimes. The soot dynamics is predicted using a Method of Moments approach with Lagrangian Interpolative Closure (MOMIC) for the fractional moments. Since no prior knowledge of the particles distribution is required, the model is generally applicable. The effect of radiation is introduced as an optically thin model. The current model accounts for the basic soot transport phenomena as transport by molecular diffusion and Thermophoretic forces. The model is first validated against experimental results for non-sooting swirling non-premixed and partially premixed flames. Next, a set of canonical premixed sooting flames are simulated, where the effect of turbulence, binary diffusivity and C/O ratio on soot formation in highly turbulent flames are studied. Finally, the model is validated against a non-premixed jet sooting flame. The effect of the flame structure on the different soot formation stages as well as the particle size distribution is described. Good results are predicted with reasonable accuracy.