Computational Fluid Dynamics - Fall 2003

1
Computational Fluid Dynamics - Fall 2003

• The syllabus
• Term project
• CFD references (Text books and papers)
• Course Tools
• Course Web Site http//twister.ou.edu/CFD2003
• Blackboard http//ou.blackboard.com
• Computing Facilities available to the class
  (accounts info will be provided)
• SOM Metlab workstations and OSCER
  (http//oscer.ou.edu) IBM Supercomputer
• Unix and Fortran Helps Consult Links at CFD
  Home page

2
Introduction Principle of Fluid Motion

• Mass Conservation
• Newtons Second of Law
• Energy Conservation
• Equation of State for Idealized Gas
• These laws are expressed in terms of
  mathematical equations, usually as partial
  differential equations.
• Most important equations the Navier-Stokes
  equations

3
Approaches for Understanding Fluid Motion

• Traditional Approaches
• Theoretical
• Experimental
• Newer Approach
• Computational - CFD emerged as the primary tool
  for engineering design, environmental modeling,
  weather prediction, among others, thanks to the
  advent of digital computers

4
Theoretical FD

• Science for finding usually analytical solutions
  of governing equations in different categories
  and studying the associated approximations /
  assumptions

h d/2,
5
Experimental FD

• Understanding fluid behavior using laboratory
  models and experiments. Important for validating
  theoretical solutions.
• E.g., Water tanks, wind tunnels

6
Computational FD

• A Science of Finding numerical solutions of
  governing equations, using high-speed digital
  computers

7
Why Computational Fluid Dynamics?

• Analytical solutions exist only for a handful of
  typically simple problems
• Much more flexible each change of
  configurations, parameters
• Can control numerical experiments and perform
  sensitivity studies, for both simple and
  complicated problems
• Can study something that is not directly
  observable (black holes).
• Computer solutions provide a more complete sets
  of data in time and space than observations of
  both real and laboratory phenomena

8
Why Computational Fluid Dynamics? - Continued

• We can perform realistic experiments on phenomena
  that are not possible to reproduce in reality,
  e.g., the weather
• Much cheaper than laboratory experiments (e.g.,
  crash test of vehicles, experimental launches of
  spacecrafts)
• May be much environment friendly (testing of
  nuclear arsenals)
• We can now use computers to DISCOVER new things
  (drugs, sub-atomic particles, storm dynamics)
  much quicker
• Computer models can predict, such as the weather.

9
An Example Case for CFD Thunderstorm
Outflo/Density Current Simulation
10
Thunderstorm Outflow in the Form of Density
Currents
11
Positive Internal Shear
g1
Negative Internal Shear
g-1
12
Positive Internal Shear
T12
g1
Negative Internal Shear
g-1
No Significant Circulation Induced by Cold Pool
13
Simulation of an Convective Squall Line in
Atmosphere
Infrared Imagery Showing Squall Line at 12 UTC
January 23, 1999.
ARPS 48 h Forecast at 6 km Resolution Shown are
the Composite Reflectivity and Mean Sea-level
Pressure.
14
Difficulties with CFD

• Typical equations of CFD are partial differential
  equations (PDE) that requires high spatial and
  temporary resolutions to represent the originally
  continuous systems such as the atmosphere
• Most physically important problems are highly
  nonlinear - true solution to the problem is often
  unknown therefore the correctness of the solution
  hard to ascertain need careful validation!
• It is often impossible to represent all relevant
  scales in a given problem - there is strong
  coupling between scales in atmospheric flows and
  most CFD problems. ENERGY TRANSFERS

15
Difficulties with CFD

• The initial condition of a given problem often
  contains significant uncertainty such as that
  of the atmosphere
• We often have to impose nonphysical boundary
  conditions.
• We often have to parameterize processes which are
  not well understood (e.g., rain formation,
  chemical reactions, turbulence).
• Often a numerical experiment raises more
  questions than providing answers!!

16
POSITIVE OUTLOOK

• New numerical schemes / algorithms
• Bigger and faster computers
• Faster network
• Better desktop computers
• Better programming tools and environment
• Better visualization tools
• Better understanding of dynamics /
  predictabilities
• etc.