CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FloEFD, FloWorks & FloTHERM (http://www.cfd-online.com/Forums/floefd-floworks-flotherm/)
-   -   FloEFD solver, staedy-state & transient (http://www.cfd-online.com/Forums/floefd-floworks-flotherm/89446-floefd-solver-staedy-state-transient.html)

Algoa June 13, 2011 07:15

FloEFD solver, staedy-state & transient
 
Hi there,
could anybody explain the difference between a steady state and a transient calculation. What criteria is the solver using to converge?

Regards
Algoa

Kevin De Smet June 13, 2011 12:51

A steady state calculation takes initial conditions and iterates on a solution based on your boundary conditions until the values stop changing much (i.e. it has converged)

You can set the initial conditions close to what you expect the results to be, or run with the defaults. It uses internally defined parameters to converge on. However you can specify engineering goals, a goal might be a mass flow rate for a pressure boundary condition, as this is one of the unknowns it needs to solve.

Transient calculations require you give precise initial conditions, and it will run for a set time and the point here is not really to converge on "one solution" but rather to see the behavior over a certain amount of time of a particular system.

Algoa June 15, 2011 03:30

Thanks for your reply Kevin,
let me ask my question with an example:
I run a steady-state calculation with mass flow rate as inlet boundary condition. As one goal pressure is set. The solver iterates until the criterion is reached/satisfied. I understand that. But does steady-state solving has any relation to time, because the inlet mass flow rate is a time depended variable?

Kevin De Smet June 19, 2011 10:59

I think it does, it all depends on your initial conditions, when solving steady-state it will take practically whatever and just take longer or shorter until things balance out. Transient analyses require precise initial conditions so you get an accurate sense of how long something takes.

Basically, if you don't really care about the time, selecting steady-state makes your life easier. It's sort of a pseudo-time.

Boris_M June 28, 2011 13:00

Hello to both of you,
last thing is not quite correct. If you define a mass flow rate in a steady-state it is not a time dependent variable. The unit of it might include the time such as kg/s but this only says that there is a constant flow of a certain amount.
A time dependent variable would be if the mass flow rate changes over the time for example if you start a pump the flow is zero at the beginning and rises over time to a "steady-state" to a point where it doesn't change anymore. Here are two examples:

"Steady-state"
If you park your car, which was running with A/C, in the sun. The motor is hot in the beginning and the interior is cool from the A/C. Now over the next hour or so the motor cools down and the interior heats up to a temperature where it reached an equilibrium with its environment and will not change over the time anymore. (Considering the sun doesn't go down and cools off at night etc.) So reaching an equilibrium is the converged state or become "steady".
For fluid flow for example the constant flow in a straight pipe doesn't change any more.

"transient"
If you want to see the changes over time such as the heating up of the car. For example you want to know after how much time the car reached from 20C the unbearable 45C. Then the calculation would show how the temperature would rise.
Or for example if you consider a pulsating flow in an engine manifold as the pistons move interchangingly.

For some flows where the flow is to turbulent and fluctuating too much you might not get a converged steady-state.
So it really depends on what you want to know, how the parameters you seek change over time (transient) or if you are interested in the final value as if the machine or whatever it is was running over a long time and the values don't change anymore.


All times are GMT -4. The time now is 17:44.