Browsing by Author "Li, Chung-Gang"

Now showing 1 - 17 of 17

An investigation of a dual-reflection phenomenon of a natural convection in a three dimensional horizontal channel without Boussinesq assumption
(2010-03-01)
Fu, Wu-Shung
;
Li, Chung-Gang
;
Tseng, Ching-Chi
;
機械工程學系
;
Department of Mechanical Engineering
An investigation of a dual-reflection phenomenon of a natural convection induced by a high temperature difference in a three dimensional horizontal channel is studied numerically. Because of a situation of high temperature difference, the Boussinesq assumption usually made in the natural convection problem is no longer available, the effects of the compressibility of fluid on all related factors must be considered. Solution methods of Roe. preconditioning and dual time stepping are appropriate for low speed compressible flow field and combined to solve the governing equations. For saving computation time, the computational process is parallel and 8 processors are used. At the aperture of the horizontal channel, the directions of fluid flows in the upper and lower regions are opposite which results in the reflection problems in the upper and lower regions induced by the large difference between the velocities of acoustic waves and fluids being completely different. The two different reflections phenomena occurring at the same aperture is conveniently named as a dual-reflection phenomenon which has never been investigated. (C) 2009 Elsevier Ltd. All rights reserved.
An investigation of a high temperature difference natural convection in a finite length channel without Bossinesq assumption
(2009-05-01)
Fu, Wu-Shung
;
Li, Chung-Gang
;
Huang, Chien-Ping
;
Huang, Jieh-Chau
;
機械工程學系
For expanding industrial applications of a natural convection in a channel under a high temperature difference, a compressible fluid flow is taken into consideration in governing equations instead of renunciation of the Bossinesq assumption. Due to the slowness of the velocity of fluid in the natural convection, the Roe scheme matching methods of preconditioning and dual-time stepping are used to solve the governing equations. In order to resolve reflections induced by acoustic waves at the boundaries of the channel, non-reflection conditions at the boundaries of the channel are derived. The results show that a more reasonable pressure distribution in the channel is first validated and average Nusselt numbers are expressed in terms of Rayleigh number for a wide range temperature difference. (C) 2009 Elsevier Ltd. All rights reserved.
An investigation of compressible forced convection in a three dimensional tapered chimney by CUDA computation platform
(2011-07-01)
Fu, Wu-Shung
;
Wang, Wei-Hsiang
;
Huang, Yun
;
Li, Chung-Gang
;
機械工程學系
Compressible forced convection in a three dimensional tapered chimney is investigated numerically. Schemes of Roe and preconditioning are adopted to resolve compressible flow problem under a low speed velocity situation. Coordinates transformation of algebraic grid generation and non-reflecting boundary condition is used to execute the coordinated transformation of the tapered duct and decrease the computation domain, respectively. Computing procedures are performed on the CUDA computation platform which is developed recently and is a very effective tool. Due to the variation of the geometry of three dimensional tapered chimney, the flow velocity increases in the central region and decreases near the corner region. Therefore, heat transfer rates are enhanced remarkably in the central region and become inferior abruptly near the corner region. The efficiency of CUDA computation platform is marvelous, and the possibility of capability of super computation of individualization may be realized by using the CUDA computation platform. (C) 2011 Elsevier Ltd. All rights reserved.
An Investigation of Compressible Turbulent Forced Convection by an Implicit Turbulence Model for Large Eddy Simulation
(2013-12-01)
Fu, Wu-Shung
;
Li, Chung-Gang
;
Tsubokura, Makoto
;
Huang, Yun
;
Domaradzki, J. A.
An investigation of compressible turbulent forced convection in a three-dimensional channel flow is studied numerically by an implicit turbulence model for large eddy simulation (LES). Because of a high temperature difference between two walls and turbulent flow, the compressibility and viscosity of fluid should be taken into consideration simultaneously. Methods of the Roe scheme, preconditioning, and dual time stepping coordinating an implicit turbulence model for LES are used for resolving the effect of the compressibility of fluid on a low speed flow field. The magnitudes of Re-tau based on the friction velocity changing from 180 to 940, with the high temperature difference of two walls of 500k are conducted. The results of the mean velocity profiles and turbulent intensities are in good agreement with the benchmark DNS data obtained by spectral codes from a low Reynolds number (Re-tau =180) to a high Reynolds number (Re-tau = 940). Besides, the larger the Re-tau is, with the exception of acquirement of larger average Nusselt number, the more drastic variation of local instantaneous Nusselt number is observed.
An investigation of natural convection in parallel square plates with a heated bottom surface by an absorbing boundary condition
(2013-01-01)
Fu, Wu-Shung
;
Wang, Wei-Hsiang
;
Li, Chung-Gang
;
Huang, Shang-Hao
;
機械工程學系
A study of natural convection in three dimensional square plates is investigated numerically. Several open boundaries co-exist in the physical domain, and a non-reflecting boundary condition applied at the aperture is no longer suitable for solving the problem of this study. An absorbing boundary condition is then adopted and modified in the solution processes. Methods of the Roe scheme, preconditioning and dual time stepping matching LSUGS scheme are used to solve a situation of a low speed compressible flow. The geometry of the physical model is the parallel square plates, and the parameter of Rayleigh number is not high, and then phenomena observed are almost symmetrical. The results of the work have good agreements with the experimental results of a previous paper. (c) 2012 Elsevier Ltd. All rights reserved.
An investigation of natural convection in parallel square plates with a heated top surface by a hybrid boundary condition
(2014-10-01)
Fu, Wu-Shung
;
Wang, Wei-Hsiang
;
Li, Chung-Gang
;
機械工程學系
;
Department of Mechanical Engineering
Natural convection in parallel squares plats with a heated top surface is investigated numerically. For solving a problem induced by different directions between the thermal diffusion and main flow streams, a hybrid boundary condition composed of the absorbing boundary condition and local one-dimensional inviscid relations (LODI) method is then adopted in artificial buffer zones. Methods of the Roe scheme, preconditioning and dual time stepping matching the Monotone Upstream-centered Schemes for Conservation Laws (MUSCL) are used to solve low speed compressible flows caused by natural convection. The results show that the hybrid boundary condition successfully prevents the rebound of heat energy from the artificial buffer zone back to the domain. The area-averaged Nusselt numbers of this study are slightly larger than those of existing experimental results, and the averaged deviation is about 10%. (C) 2014 Elsevier Masson SAS. All rights reserved.
An investigation of natural convection in parallel square plates with a heated top surface by a hybrid boundary condition
(2014-10-01)
Fu, Wu-Shung
;
Wang, Wei-Hsiang
;
Li, Chung-Gang
;
機械工程學系
;
Department of Mechanical Engineering
Natural convection in parallel squares plats with a heated top surface is investigated numerically. For solving a problem induced by different directions between the thermal diffusion and main flow streams, a hybrid boundary condition composed of the absorbing boundary condition and local one-dimensional inviscid relations (LODI) method is then adopted in artificial buffer zones. Methods of the Roe scheme, preconditioning and dual time stepping matching the Monotone Upstream-centered Schemes for Conservation Laws (MUSCL) are used to solve low speed compressible flows caused by natural convection. The results show that the hybrid boundary condition successfully prevents the rebound of heat energy from the artificial buffer zone back to the domain. The area-averaged Nusselt numbers of this study are slightly larger than those of existing experimental results, and the averaged deviation is about 10%. (C) 2014 Elsevier Masson SAS. All rights reserved.
An investigation of the unstable phenomena of natural convection in parallel square plates by multi-GPU implementation
(2017)
Fu, Wu-Shung
;
Wang, Wei-Hsiang
;
Li, Chung-Gang
;
Tsubokura, Makoto
;
機械工程學系
Unstable phenomena induced by natural convection of parallel square plates are investigated numerically. The geometry belongs to the open boundary problem, and the compressibility of fluids is taken into consideration. The absorbing boundary condition and modified local one-dimensional inviscid relation method are adopted for open boundaries. Numerical methods of the Roe scheme, preconditioning, and dual time stepping matching the data-parallel lower-upper relaxation method are coordinated with multi-GPU implementation. Unstable phenomena of natural convection caused by the buoyancy force can be observed. Present results have good agreement with the experimental results in the overlap range of the existing study.
An investigation of unsteady flow reversal of natural convection in vertical parallel plates by the modified local one-dimensional inviscid relations method
(2015-07-01)
Fu, Wu-Shung
;
Huang, Shang-Hao
;
Li, Chung-Gang
;
機械工程學系
;
Department of Mechanical Engineering
An unsteady flow reversal of natural convection in vertical parallel plates with an asymmetrically heated wall is investigated numerically. For clarifying the occurrence of flow reversal, the width of the plates is regarded as a variable, and the non-reflecting boundary condition is adopted at apertures. The methods of the Roe scheme, preconditioning and dual time stepping matching LUSGS are siinultaneously used to solve governing equations which are available for a low speed compressible flow problem. The flow reversal is observed under broad width situations and absent under narrow width situations. Distributions of the pressure differences at the apertures are revealed to indicate the occurrence of flow reversal. The present Nusselt numbers, which mean time and area-averaged ones, display relatively good agreement with those of the existing works. (C) 2015 Elsevier Ltd. All rights reserved.
Applying Compute Unified Device Architecture to Investigate Gaseous Discharge from High-Pressure Vessel
(2016-01)
Fu, Wu-Shung
;
Huang, Kun-Rong
;
Wang, Wei-Hsiang
;
Li, Chung-Gang
;
機械工程學系
The aim of this study is to investigate gases discharged from a high-pressure vessel numerically. To simulate this subject more realistically, the viscosity and compressibility of the gas are taken into consideration simultaneously. The methods of the Roe scheme, preconditioning, and dual time-stepping matching the lower-upper symmetric-Gauss-Seidel method are adopted to solve compressible flow problems during gaseous discharge processes. The nonreflecting boundary condition is used to prevent flowfields from being polluted by the reflection of the pressure wave induced by the compressible flow at the boundary. Computing procedures are performed on the compute unified device architecture computation platform, which was recently developed and is a highly effective technology for accelerating computational speed. Results show that the mass flow rate of this work is consistent with the existing experimental work. Because of a sudden expansion at a small opening, the phenomena of an alternating variation of the pressures of gases, rapid decrements of the temperature of gases, and a quick acceleration of the velocities of gases are remarkably observed in the mainstream direction. The ratio of the thrust caused by the gases released to the reaction force is less than 1 because of the dissipation of entropy generation.
Compressible direct numerical simulation with a hybrid boundary condition of transitional phenomena in natural convection
(2015-11-01)
Li, Chung-Gang
;
Tsubokura, Makoto
;
Fu, Wu-Shung
;
Jansson, Niclas
;
Wang, Wei-Hsiang
We investigate the phenomena involved in the transition from laminar to turbulent flow in natural convection through a channel using compressible direct numerical simulation (DNS). The Roe numerical scheme with preconditioning and dual time stepping is employed to handle slow natural convection flows with large temperature differences and variable densities. A hybrid boundary condition that combines absorbing and non-reflecting boundary conditions is applied at the inlet and outlet, which does not require a priori knowledge of the flow rate. No empirically derived turbulence parameters are required in the DNS code. The numerical predictions are qualitatively consistent with published experimental data and the transition point can be accurately captured. Our results show that the transitional phenomena can be naturally induced by the interactions between the buoyancy force and shear stress without prescribing any fluctuations at the inlet. We submit that the numerical scheme and model configuration developed in this study has the potential to be a benchmark case for evaluating the accuracy of other turbulence models. (C) 2015 Elsevier Ltd. All rights reserved.
Direct numerical simulation of film cooling with a fan-shaped hole under low Reynolds number conditions
(2018-08-01)
Fu, Wu-Shung
;
Chao, Wei-Siang
;
Tsubokura, Makoto
;
Li, Chung-Gang
;
Wang, Wei-Hsiang
Shaped film cooling holes have several features that can greatly improve film cooling effectiveness, and it has been studied and utilized in gas turbine engines for decades. Few studies, however, have reported the effects of low mainstream Reynolds number on shaped film cooling holes. In this study, the effects of mainstream Reynolds number on film cooling with a fan shaped hole are studied by utilizing direct numerical simulation (DNS). In addition, the compressibility and the viscosity of the working fluid are simultaneously considered, and the non-reflecting and absorbing boundary conditions are adopted at the exit of the main channel. The methods of the Roe scheme, preconditioning, and dual time stepping are employed together to solve the governing equations of a low-speed compressible flow problem. This study considers the mainstream Reynolds numbers of Re-D = 480 and 3200 with 0% and 5% turbulence intensity in the mainstream. Results reveal that the coolant jet penetrates into the mainstream with a mainstream Reynolds number of 480. However, at the higher Reynolds number, the coolant jet develops along the wall and results in better film cooling effectiveness. In addition, special attention is paid to the structures of the vortices developed from the crossflow. Hairpin vortices become smaller at higher mainstream Reynolds numbers. On the contrary, horseshoe vortices appear when the mainstream Reynolds number is increased. A detailed comparison of the vortices is presented in this study. (C) 2018 Elsevier Ltd. All rights reserved.
Enhancement of forced convection heat transfer in a three-dimensional laminar channel flow with insertion of a moving block
(2010-09-01)
Fu, Wu-Shung
;
Huang, Jieh-Chau
;
Li, Chung-Gang
;
機械工程學系
;
Department of Mechanical Engineering
Enhancement of heat transfer rate of a heat surface in a three-dimensional channel flow with insertion of a moving slender block is investigated numerically. A slender block is installed along the direction of the channel flow, and the movement of the slender block is in periodic motion and transverse to the channel flow. Due to the interaction between the channel flow and slender block moving, a large circulation vertical to the channel flow is induced and follows the moving slender block closely. The behaviors of the slender block and circulation destroy and suppress the velocity and thermal boundary layers on the heat surface periodically. These phenomena first observed cause the heat transfer rate of the heat surface to be enhanced especially in the downstream of the heat surface. The maximum magnitude of enhancement in this study is about 50%. (C) 2010 Elsevier Ltd. All rights reserved.
Estimation of turbulent natural convection in horizontal parallel plates by the Q criterion
(2013-07-01)
Fu, Wu-Shung
;
Lai, Yu-Chih
;
Li, Chung-Gang
;
機械工程學系
;
Department of Mechanical Engineering
An investigation of estimation of turbulent structures of natural convection in horizontal parallel plates by the Q criterion is studied numerically. It is difficult to adopt a definite criterion like the critical Reynolds number adopted in forced convection to distinguish laminar and turbulent flows of natural convection in horizontal parallel plates because the angle between directions of the buoyancy force and main stream flow is changed in the whole process of the flowing. The Q criterion is then used to make up for a deficiency of the definite criterion and examines mechanisms of the occurrence of rotation behaviors which are often regarded as turbulent structures in detail. Methods of the Roe scheme, preconditioning and dual time stepping matching the method of LUSGS are simultaneously used to solve a low speed compressible flow. As well, non-reflecting boundary conditions are held at apertures to prevent the reflection caused by pressure waves. Under a low height situation, the shear force is dominant and then a few rotation behaviors are observed near the central region where flow directions are changed. Oppositely, the space is almost filled with active rotation behaviors because natural convection fully develops in a high height situation. (C) 2013 Elsevier Ltd. All rights reserved.
Flow downward penetration of vertical parallel plates natural convection with an asymmetrically heated wall
(2016-05)
Fu, Wu-Shung
;
Chao, Wei-Siang
;
Peng, Tzu-En
;
Li, Chung-Gang
;
機械工程學系
Phenomena of the flow downward penetration in vertical parallel plates natural convection with an asymmetrically heated wall are investigated numerically. For clarifying the occurrence of flow downward penetration, the width of parallel plates is further broadened and broader than that investigated in the previous study in which the flow reversal was studied in detail. In order to simulate realistically, the compressibility and the viscosity of the working fluid are simultaneously considered, and the non-reflecting boundary is adopted at apertures of plates. The methods of the Roe scheme, preconditioning, and dual time stepping matching LUSGS are coordinated to solve governing equations of a low-speed compressible flow problem. The flow downward penetration, in which fluids are sucked from surroundings outside the outlet to flow into the parallel plates and through the inlet flow out of parallel plates, is composed of a few fluids and firstly observed. Also, two flow reversal caused by the buoyancy force appear in the upper left region, and one flow reversal induced by the flow downward penetration appears in the low right region. The usage of the existing results and present study, an equivalent ratio of the width to the length of parallel plates, of which the heat transfer rate is the same as the single vertical heated plate, can be expediently obtained. In this study, the ratio is about 0.6. (C) 2016 Elsevier Ltd. All rights reserved.
Investigation of boundary layer thickness and turbulence intensity on film cooling with a fan-shaped hole by direct numerical simulation
(2018-08-01)
Fu, Wu-Shung
;
Chao, Wei-Siang
;
Tsubokura, Makoto
;
Li, Chung-Gang
;
Wang, Wei-Hsiang
Effects of the mainstream boundary layer thickness and the turbulence intensity on film cooling under low Reynolds number conditions are studied in this work by the direct numerical simulation (DNS). In other to solve low-speed compressible flow problems, several methods of Roe scheme, preconditioning, dual time stepping, and LUSGS are adopted to solve governing equations. Results reveal that a horseshoe vortex appears with a thicker mainstream boundary layer, and thus the lateral coverage of the coolant fluid has increased significantly. In addition, the existence of turbulence intensity eliminates the blow-off phenomenon, which happens in a thin mainstream boundary layer condition and enhances the film cooling effectiveness.
Roe scheme with preconditioning method for large eddy simulation of compressible turbulent channel flow
(2009-11-20)
Fu, Wu-Shung
;
Li, Chung-Gang
;
Lin, Wei-Fong
;
Chen, Yu-Hsu
;
機械工程學系
Numerical investigations of fully compressible turbulent channel flows are conducted. Two different speeds of Mach numbers of 0.005 and 0.5 are separately taken into consideration to study mechanisms of compressible turbulent flows and increments of temperatures of working fluids. For computing the compressible low-speed turbulent flow efficiently, methods of Roe and preconditioning matching LES methods are adopted. In addition, usage of dual time stepping and curved linear coordinates transformation skills to stabilize transient situations and increase computing grids near the channel walls are performed. As for an assignation of initial conditions, random velocity fluctuations are imposed on the mean velocity distribution. The results show that the mean velocity distributions, root-mean-square values of fluctuating velocities and mean Reynolds stresses have good agreements with the existing results. The increments of temperatures of working fluids are remarkable for a Mach number of 0.5 and are also well consistent with the existing results. Copyright (C) 2008 John Wiley & Sons, Ltd.