CTR Seminars Archive 2013
Date: Friday December 13, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Benedetta Franzelli, CTR Postdoctoral Fellow
Title: Understanding and modeling of chemical-related phenomena in purely gaseous and multi-phase reactive flows
Abstract: The maturity of propulsion technologies for aerospace and land transport means that the performance gains and emission reduction can only be achieved by a near-perfect control of all physical phenomena involved during combustion. Increasingly, high-performance DNS and LES calculations are performed to reproduce and understand the complex chemical-related phenomena (flame stabilisation, thermo-acoustic instabilities, extinction and auto-ignition, pollutant emission…). This requires the development of models that allow for a detailed description of the phenomena of interest at a reasonable computational cost.
Detailed kinetic schemes exist for most hydrocarbons but their use in turbulent combustion simulations is limited due to their high computational cost. Reduced kinetic mechanisms and tabulation methods are initially developed on laminar configurations to overcome this problem for both purely gaseous and two-phase flames. The validity of such simplified descriptions is then evaluated on DNS and LES of turbulent configurations. Their ability to predict both the structure and the dynamics of the flame as well as pollutant emissions, such as carbon mono-oxide, is investigated. The construction of a simplified chemical method for DNS and LES of turbulent 3-D configurations requires a compromise between the computational cost and quality of the results. This compromise can be assessed by analyzing the behavior of chemical methods in simplified one-dimensional laminar flames.
Date: Wednesday December 11, 2013
Time: 3:30pm
Location: CTR Conference Room
Speaker: Dr. Jeroen A.S. Witteveen, Scientific Staff Member, Center for Mathematics and Computer Science (CWI), Amsterdam, The Netherlands
Title: Stochastic Power Flow using Gauss Quadrature Sparse Grids
Abstract: The growing capacity of renewable energy sources in combination with more extreme weather fluctuations lead to increasing variability in the electricity production. However, supply and demand of electricity have to be in equilibrium at all times to avoid frequency instabilities in the electrical grid and power blackouts. Nowadays, the short-term operation and long-term grid extension planning is performed based on deterministic scenarios and purely historical data. Stochastic power flow simulations are therefore needed to reduce computational time by focusing on the most probable scenarios and to incorporate possible new future loading conditions. It is known that the non-standard data-driven distributions, which are collected in this case, can lead to negative quadrature weights and resulting negative variances for Clenshaw-Curtis sparse grids. On the other hand, Gauss quadrature rules always have positive weights for any distribution, but they are expensive on a sparse grid, because their abscissas are not nested. Therefore, we construct here nested subsets of Gauss quadrature points to obtain optimal accuracy on the highest level as well as efficient sparse grids and error estimates on the lower levels. Numerical results confirm that these Gauss quadrature sparse grids are more accurate than Clenshaw-Curtis points, also on the lower levels.
Date: Friday Jul 12, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Prof. Tianfeng Lu, University of Connecticut, Dept. of Mechanical Engineering
Title: Accommodating Complex Chemistry in Large-Scale Flame Simulations
Abstract: Detailed chemistry is an integral component of predictive flame simulations. However, detailed chemical kinetics of realistic fuels can be highly complex due to the large number of species and reactions and the severe chemical stiffness induced by highly reactive radicals. Challenges to accommodate detailed chemistry in large-scale flame simulations are two-fold. First, the large mechanisms must be reduced before they can be applied to multi-dimensional flame simulations, resulting in the need of rigorous and efficient methods for mechanism reduction and stiffness removal. Second, 3-D simulations may involve complex turbulent-flame interactions and output massive simulation data, e.g. in hundreds of terabytes. As such, systematic computational diagnostics are required to extract salient information from the massive simulation results and to understand the complex flame behaviors and the underlying physicochemical processes. For mechanism reduction, a suite of algorithms including directed relation graph (DRG), analytically solved quasi steady state approximations, and dynamic chemical stiffness removal were developed to systematically obtain compact, accurate and non-stiff mechanisms from detailed mechanisms with hundreds or thousands of species, for a variety of fuels ranging from methane to surrogates of gasoline, kerosene, diesel and biodiesel. For computational flame diagnostics, a method of chemical explosive mode analysis (CEMA) was developed to systematically identify limit flame phenomena, such as ignition, extinction and onset of flame instabilities, and other critical flame features, such as premixed flame fronts and non-premixed flame kernels, in complex flame configurations at both laminar and turbulent conditions. In the presentation, the usage of CEMA for flame diagnostics will be demonstrated with Sandia’s recent direct numerical simulations (DNS) of lifted jet flames, homogeneous charge compression ignition (HCCI) combustion, and premixed and non-premixed temporal jet flames.
Date: Friday Apr 12, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Jean-Pierre Hickey, Center for Turbulence Research, Postdoctoral Fellow
Title: Coherent Structures in the Evolution of Sub- and Supersonic Planar Wakes
Abstract: Wakes are constitutive components of engineering, aeronautical and geophysical flows. Despite the canonical nature of wakes, many fundamental questions surrounding this simple flow remain unanswered. The present work revisits the nature of archetypal planar splitter-plate wakes in the sub- and supersonic regimes, more specifically addressing the effects of the imbedded coherent structures on the evolution of the flow. In this presentation, three topics will be discussed. Firstly, we will argue that the far wakes maintain a memory of their origin, and the resulting multiplicity of the self-similar states can be tied to a plurality of the large-scale, far wake structural organizations. Secondly, building on linear theory and vortex dynamics, we will infer the concomitant Mach number effects on the emerging structures during the transition of high speed wakes. Finally, we will use critical point theory to explain the possibility of a breakdown mechanism in the inclined vortex ribs of the high speed transitioning wake.
Date: Friday Mar 29, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Florian Kumer, Fluid Dynamics, Darmstadt University of Technology
Title: Discontinuous Galerkin methods for multi-phase flows
Abstract: For the numerical simulation of multi-phase flows, a sharp representation of the fluid interface is a challenging task, especially in the context of thigh-order methods, like the discontinuous Galerkin (DG) method. In order to preserve the hp-convergence property of the DG method, not only an accurate description of the position of the fluid interface is necessary. Furthermore, a special numerical treatment for the jumps and kinks which occur in the pressure and velocity field of the flow is required. Our approach is to track the interface position by a high order level-set method. The discontinuities in the solution will be represented by an adaption of the numerical approximation space. Within the talk, an overview about my research project for developing such a method will be given, and how these attempts are linked to the broader research activities at my home institution in Germany. I will discuss what has been achieved so far, and give an outline on my research activities within the next two years at the CTR.
Date: Friday Mar 15, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Kourosh Shoele, UC San Diego and Re-Vision Consulting LLC
Title: Fluid Interaction with Structures: From Fish Fins to Hydrokinetic Devices
Abstract: Studying the interaction between fluid and structure is a fundamental step in understanding the underpinnings of many engineering and physical phenomena, from energy harvesting to biolocomotion of insects, birds and fishes. The complex nature of these interactions makes the design of computational, experimental, and analytical techniques for modeling such problems challenging. Here I discuss new procedures, both in potential flow and viscous flow, for studying the interactions of a flexible structure with a flow. In particular, I will focus on two particular phenomena, the flow interaction with skeleton-reinforced fish fins and the extraction of ocean energy through oscillating systems. 1) Fins of bony fishes are characterized by a skeleton-reinforced membrane with a soft collagen membrane strengthened by embedded flexible rays. Our results illustrate that the fish’s capacity to control the motion of each individual ray, as well as the anisotropic deformability of the fin, are essential to high propulsion performance. By drawing connections with a set of phenomena, we demonstrate that this structural design is a recurring feature in nature. 2) The ocean offers numerous resources for renewable energy, from winds, waves, currents and tides. The harvesting of ocean energy is a relatively new and emerging research area. I discuss our recent efforts to develop a unified multilevel package for analysis, prediction and control of different floating power systems using Bond graph object-oriented modeling approach. In particular, the results illustrate the benefits of active controlling of wave converters such as a point absorber wave converter and an oscillatory water column (OWC) device.
Date: Friday Mar 8, 2013
Time: 4:00pm
Location: BLDG 530 ROOM 127
Speaker: André Thess, Institute of Thermodynamics and Fluid Mechanics, Ilmenau University of Technology, Germany
Title: Electromagnetic Flow Measurement in Metallurgy, Crystal Growth and Energy Conversion
Abstract: The measurement of velocity in liquid metals, semiconductor melts and molten salts is a notoriously difficult problem because these materials are often opaque, hot and aggressive. Hence the development of reliable contactless velocity measurement methods is a challenge with considerable practical impact. Electromagnetic flow measurement is a promising technique to solve this problem. In the first part of the presentation we give a brief overview of the history and new developments of electromagnetic flow measurement starting from Michael Faradays experiments in 1832 to new techniques like eddy current flowmeters, contact less inductive tomography and Lorentz force velocimetry. In the second part of the presentation we discuss several examples that demonstrate the important role of computational fluid dynamics and computational magnetohydrodynamic in the development of such flowmeters. We finally discuss some possible new applications including applications in solar thermal power plants.
Date: Friday Mar 1, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Sabre Bougrine, Center for Turbulence Research, Postdoctoral Fellow
Title: 0-Dimensional Modeling of the Combustion of Alternative Fuels in Spark-Ignition Engines
Abstract: The work aimed at proposing a physical combustion chamber system simulation model able to accurately represent the different phenomena associated to all common technologies (GDI, EGR, downsizing, VCR, CAI, etc.) and to the use of alternative fuels (CNG, hydrogen, gasoline, ethanol and liquid or gaseous hydrocarbon mixtures). It has resulted in a new 0D combustion model, CFM1D-TC (Tabulated Chemistry), derived from the 3D CFD ECFM model and integrating several new approaches to describe turbulent premixed flame propagation, auto-ignition and pollutant formation processes.
Date: Friday Feb 15, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Prof. Julian Hunt, University College London, Trinity College Cambridge, TU Delft
Title: Interfacial layers in high Re turbulence-physics and simulations
Abstract: Where turbulent flows with or without mean shear are adjacent to regions of very weak turbulence, such as boundary layers wakes, jets, and plumes, or convective layers , quasi-continuous interfacial layers tend to form that separate the turbulent and non-turbulent regions. Measurements and computations show that the mean and variance profiles of velocity and scalars have common global features, such as the outward boundary entrainment velocity’ and discontinuities of the Reynolds stresses .These features vary depending on the strength of the shear . Microscale vortices are common to most interfacial layers. These results can be explained in terms of key mechanisms, including instabilities, shear sheltering, nibbling/engulfment, and energy transfer dynamics in the vicinity of the interfaces. Improvements in parameterisations and in under -resolved simulations are emerging. Computations by Ishihara and Kaneda on the Earth Simulator at Rlamda greater than 1000 and laboratory turbulence at Cambridge at Rlamda up to 450 show how similar thin-layers occur intermittently in the interior of turbulent flows. But there are some important differences. Detailed analysis of the data and a rapid distortion model shows how the eddies impact on the layers leads to a net down scale transport, peak dissipation and microscale velocities of the order of the RMS velocities in the layers. Some of the statistics are consistent with Kolmogorov theory.
Date: Friday Feb 8, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Prof. Charles Tinney, The Univ. of Texas at Austin, Aerospace Engineering and Engineering Mechanics
Title: Cumulative nonlinear distortion of acoustic waves produced by high-speed jet flows
Abstract: In supersonic jets, an extended line of distributed sources act in compliance to generate and propagate noise in a complex manner to an observer far away from the jet. Nonlinear distortion of the acoustic waveform is often considered a prerequisite to understanding this propagation process. This is driven by one's capture of 'shock-type' structures or so called 'N-wave' type signatures in the pressure waveform, which are generally attributed to nonlinear wave steepening. Likewise, imperfect collapse of the spectra between an observer signal and a prediction, formed from a linear rescaling of the closer signal (using 1/r2 dependence), is also believed to be caused by nonlinear distortion. Unfortunately, these observations are most often made using measurements acquired in a range-restricted environment where changes to the waveform, due to cumulative nonlinear distortion, are too small to be accurately quantified, and/or without prior knowledge of the sound propagation path. The current work aims to fill this gap by exploring the acoustic field produced by an unheated, perfectly expanded, Mach 3 jet in a laboratory-scale environment. This talk focuses on a time-averaged approach to understanding the degree of non-linearity in the acoustic waveform at several far-field observer positions. Various "off-the-shelf" indicators are explored including a new model for predicting the presence (or lack there of) of cumulative non-linear waveform distortions in the signal.
Date: Friday Jan 18, 2013
Time: 4:00pm
Location: CTR Conference Room
Speaker: Dr. Julien Bodart, Center for Turbulence Research, Postdoctoral Fellow
Title: Wall-modeling for large eddy simulations in complex geometries
Abstract: Accurate predictions of the flow field over complex aerodynamic configurations, such as high lift devices is of particular interest for airframe noise predictions as well as wing design with or without active flow control. In particular, computing the flow at high angle of attack and in particular predicting the maximum lift coefficient remains a challenge using RANS solvers. We discuss how large eddy simulation can be used in this context, together with wall-modeling to reduce the grid requirements. I will specifically address two challenges wall-modeling has to face when dealing with complex geometries: i) its implementation in massively parallel, unstructured solvers and ii) its ability to predict transitional flows. The wall-model methodology implemented in the solver CharLES^X relies on flow sensors, to activate the wall model only in turbulent regions of the boundary layers. I will first discuss results obtained for a canonical (H-type) transitional boundary layer, and for the flow around McDonnell-Douglas 30P/30N airfoil, at the realistic Reynolds number (based on the stowed chord) of Re_c=9.10^6.