Bulletin of the American Physical Society
66th Annual Meeting of the APS Division of Fluid Dynamics
Volume 58, Number 18
Sunday–Tuesday, November 24–26, 2013; Pittsburgh, Pennsylvania
Session L5: CFD VI |
Hide Abstracts |
Chair: Meng-Sing Liou, NASA Room: 327 |
Monday, November 25, 2013 3:35PM - 3:48PM |
L5.00001: On Numerical Heating Meng-Sing Liou The development of computational fluid dynamics over the last few decades has yielded enormous successes and capabilities that are being routinely employed today; however there remain some open problems to be properly resolved. One example is the so-called overheating problem, which can arise in two very different scenarios, from either colliding or receding streams. Common in both is a localized, numerically over-predicted temperature. Von Neumann reported the former, a compressive overheating, nearly 70 years ago and numerically smeared the temperature peak by introducing artificial diffusion. However, the latter is unphysical in an expansive (rarefying) situation; it still dogs every method known to the author. We will present a study aiming at resolving this overheating problem and we find that: (1) the entropy increase is one-to-one linked to the increase in the temperature rise and (2) the overheating is inevitable in the current computational fluid dynamics framework in practice. Finally we will show a simple hybrid method that fundamentally cures the overheating problem in a rarefying flow, but also retains the property of accurate shock capturing. Moreover, this remedy (enhancement of current numerical methods) can be included easily in the present Eulerian codes. [Preview Abstract] |
Monday, November 25, 2013 3:48PM - 4:01PM |
L5.00002: Numerical Investigation of Conjugate Heat Transfer in a Channel with a Growing Deposit Layer Hongying Li, Yitfatt Yap, Jing Lou, John Chai The working fluid carries particles flowing in channels is widely encountered in many engineering applications such as oil and gas pipes and heat exchangers. These particles have a tendency to deposit onto the wall of the channels, form a deposit layer. This additional growing and increasingly thicker deposit layer, normally is of a lower thermal conductivity. In the system with heat transfer involved, such deposit layer introduces extra thermal resistance and consequently leads to the lower the heat transfer performance of the system. Besides, the deposit layer reduces flow cross sectional area of the channel and directly responsible for inducing a larger pressure drop. As such, a good understanding of the conjugate heat transfer coupling the evolving deposit layer and fluid flow is important. This numerical study is undertaken to fill in some of the gaps in this respect. Here, we consider conjugate heat transfer in a channel with a deposit layer gradually growing on the wall. The problem is governed by conservation equations for mass, momentum, species and energy, coupled with the appropriate interfacial condition at the depositing front separating the fluid from the deposit. This is a moving boundary problem as the front evolves over time. The depositing front is captured using the level-set method in this study. Numerical solution is performed on a fixed mesh using the finite volume method. A detailed parametric study quantifying the effect of the growing deposit layer on the heat transfer performance is performed. [Preview Abstract] |
Monday, November 25, 2013 4:01PM - 4:14PM |
L5.00003: Upscale and downscale energy transfer in turbulent open channel flow Salvatore Lovecchio, Alfredo Soldati Heat and mass transfer phenomena in free-surface turbulence are of great importance in a wide range of geophysical/environmental situations. Examples include $CO_2$ transfer across the ocean surface or the transport of organic species. These phenomena are controlled by the dynamics of free-surface turbulent structures, which are known to give rise to transport of energy among the flow scales. In this study we use Direct Numerical Simulation to analyze such energy transfer in turbulent channel flow with a free surface. Our results suggest that the inhomogeneity, inherently present in near-wall and free-surface turbulence, generates energy fluxes that correspond to a spatial redistribution of turbulent kinetic energy within the flow. We show that the energy transfer near the boundaries is significantly different from that in the bulk flow, where the behaviour is more homogeneous and isotropic. This is due to an increased energy backscatter from small to large flow scales. We also show that regions of direct (downscale) and inverse (upscale) energy transfer can be associated to the coherent structures of the flow. [Preview Abstract] |
Monday, November 25, 2013 4:14PM - 4:27PM |
L5.00004: Experimental Validation Dataset for CFD Simulations of Buoyancy Opposed Convection Jeff Harris, Blake Lance, Barton Smith New experiments in the Rotatable Buoyancy Tunnel are described. This unique facility was built specifically for computational fluid dynamics (CFD) validation experiments in natural, forced, and mixed convection in both buoyancy aided or opposed scenarios. The tunnel features clear walls for non-intrusive optical measurements, a heated wall (controlled to isothermal or constant flux conditions), and the capability to invert without changing the inlet or as-built dimensions. The wall temperature and inlet temperature are measured, along with the inlet velocity and turbulence profiles, to define simulation boundary conditions. The experiment includes acquisition of particle image velocimetry data at several streamwise locations in the boundary layer along the heated plate. Heat flux at those locations is also measured. The flow consists of natural convection driving air upwards combined with forced convection (fan driven) drawing air down. A RANS CFD simulation for this scenario is presented, with a comparison of several models' computed boundary layer flow, heat flux, and pressure drop to the measured values of the same. [Preview Abstract] |
Monday, November 25, 2013 4:27PM - 4:40PM |
L5.00005: Computational Fluid Dynamics Uncertainty Analysis applied to Heat Transfer over a Flat Plate Curtis Groves, Marcel Ilie, Paul Schallhorn There have been few discussions on using Computational Fluid Dynamics (CFD) without experimental validation. Pairing experimental data, uncertainty analysis, and analytical predictions provides a comprehensive approach to verification and is the current state of the art. With pressed budgets, collecting experimental data is rare or non-existent. This paper investigates and proposes a method to perform CFD uncertainty analysis only from computational data. The method uses current CFD uncertainty techniques coupled with the Student-T distribution to predict the heat transfer coefficient over a flat plate. The inputs to the CFD model are varied from a specified tolerance or bias error and the difference in the results are used to estimate the uncertainty. The variation in each input is ranked from least to greatest to determine the order of importance. The results are compared to heat transfer correlations and conclusions drawn about the feasibility of using CFD without experimental data. The results provide a tactic to analytically estimate the uncertainty in a CFD model when experimental data is unavailable. [Preview Abstract] |
Monday, November 25, 2013 4:40PM - 4:53PM |
L5.00006: ABSTRACT WITHDRAWN |
Monday, November 25, 2013 4:53PM - 5:06PM |
L5.00007: Fast geometric sensitivity analysis in hemodynamic simulations using a machine learning approach Sethuraman Sankaran, Leo Grady, Charles Taylor In the cardiovascular system, blood flow rate, velocities and blood pressure are governed by the Navier-Stokes equations. Inputs to the system such as (a) geometry of arterial tree, (b) clinically measured blood pressure and viscosity, (c) boundary resistances, among others, are typically uncertain. Due to a large number of such parameters, there is a need to efficiently quantify uncertainty in solution fields in this multi-parameter space. We use a machine learning approach to approximate the simulation-based solution. Using an offline database of pre-computed solutions, we compute a map (rule) from the features to solution fields. This is coupled to an adaptive stochastic collocation method to quantify uncertainties in input parameters. We achieve significant speed-up ($\sim$1000 fold) by approximating the simulation-based solution using a machine learning predictor. Bagged decision tree was found to be the best predictor among many candidate regressors (correlation coefficient $\sim$0.92). The sensitivities obtained using machine learning approach has a correlation coefficient of 0.91 with those obtained using finite element simulations. We also calculated and ranked the impact of different inputs such as problem geometry, and clinical parameters. We observed that the impact of geometry supersedes the impact of other variables. Mostly, segments with significant disease in the larger arteries had the highest sensitivities. We were able to localize sensitive regions in long segments with a focal disease using a multi-resolution approach. [Preview Abstract] |
Monday, November 25, 2013 5:06PM - 5:19PM |
L5.00008: ABSTRACT WITHDRAWN |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700