Simulations of fluid structure interactions

Recent advancement in computing technology has given us a way to study fluid problems numerically, this area of study is called Computational Fluid Dynamics (CFD). Early CFD is normally used to solve problem where only the dynamics of the fluid which is concerned, but more and more people are curren...

Full description

Bibliographic Details
Main Author: Koe, Han Beng.
Other Authors: Jorg Uwe Schluter
Format: Final Year Project (FYP)
Language:English
Published: 2011
Subjects:
Online Access:http://hdl.handle.net/10356/45935
Description
Summary:Recent advancement in computing technology has given us a way to study fluid problems numerically, this area of study is called Computational Fluid Dynamics (CFD). Early CFD is normally used to solve problem where only the dynamics of the fluid which is concerned, but more and more people are currently looking into Fluid Structure Interaction (FSI) since many engineering problem is essentially the interaction of the body and uid. In FSI, the boundary of the problem usually is not fixed, this will introduce problem to the current CFD codes which normally employ body-fitted mesh. In this project, another approach to simulate FSI problems is to use the Immersed Boundary Method (IBM). With IBM, a simple cartesian mesh can be used since IBM handles the existing of a body di erently. For the purpose of this project, the FSI problem that we are going to investigate is an elastically mounted cylinder undergoing Vortex-Induced Vibration (VIV). Since the existing code is only capable of handling specified motion of the body, some modfication was done on the code so that it is capable to handle spring and damper system. The fluid solver which is based on pressure correction technique will be coupled with the dynamical equation (spring and damper system) which is solved using 4th order Runge Kutta method. Some validation of the code is done by performing simulations on the flow around a static circular cylinder. It is found that some differences with the available data, this is mainly because of the laminar inflow which we used. Nevertheless, the existing code is capable of handling the simulation pretty well. Next VIV simulation is performed at Re = 8000, relatively low mass ratio m* is used in the simulation since real VIV problem is usually in this region. The reduced velocity U*, damping ratio and mass ratio m* which define the spring and damper system will be varied to study how they will affect the response of the cylinder, specifically how they affect the maximum oscillation amplitude and oscillation frequency is studied.