[Shahr]

Hello,

I am trying to perform a transient simulation of a standard laminar flow around sphere benchmark problem using CFX 14 from Workbench, but get strange pressure behaviour at second time step (AnimationFrame000003).

Here is a link to transient results:

As you can see pressure in the right side (outlet) is higher than left side (inlet).

The geometry is made according to Turek's benchmark found at:
http://www.mathematik.tu-dortmund.de...rTurek1996.pdf

non-dimensional time step size is 0.1, and Re=100. I cannot think of an explanation for this pressure behaviour. Does anybody have any idea why this happens?

P.S. I used a sphere instead of cylinder positioned at the same location.

[ghorrocks]

Have a think about your initial conditions and what would happen. I am guessing the initial condition is either no flow of just Vinf to the right. So if a sphere magically appeared in this flow then you would expect pressure waves to bounce around until things settle down a bit - and this is exactly what you are seeing.

[Shahr]

Thanks Mr. Horrocks,

I prescribed the inlet velocity and a regression polynominal of pipe flow without the obstacle as the initial condition for the fluid domain. So what you said totally makes sense.

You know, I have to prescribe both velocity and pressure field for initialisation. In the simple case I can use the steady-state results for this purpose. But, what should I do in case this sphere is to be considered as rigid body. I mean, there will be no steady-state result for a mobile rigid body sphere. What should I choose as initial condition then?

[ghorrocks]

It depends on what you are tryign to model. If you only care about the periodic transient pattern (that is shedding vorticies) then don't worry about the initial transient and just keep going. Run it until it achieves a repeating pattern.

If you are interested in the initial startup then you need to specify a start up which is physically possible. Either the flow is initially stationary and a pressure is generated which starts the flow, or the flow is already going and the sphere is swung down into the flow. You cannot have a sphere which magically appears in the middle of a flow as it is not physically possible.

[Shahr]

As you know initial condition should be defined for a transient simulation in cfx like below:


When I input inlet velocity as initial condition to get a fully-developed flow, cfx can't solve with static pressure option set to automatic and returns this error:
| FATAL ERROR : |
| |
| Initial values are required for all variables in TRANSIENT runs. |
| In this simulation, no initial value was set for |
| |
| Variable : Pressure |
| Domain : Fluid

If I use the pressure drop from simple pipe flow, I get pressure waves. This pressure wave created in my simulation domain affects the dynamics of the mobile particle and I don't want to have it. and I cannot use the steady-states results because it doesn't describe the pressure/velocity of a movable particle.

[ghorrocks]

As I said in my previous post, then you need some physically possible way of starting the flow. You could start the velocity from zero and ramp it up, or maybe swing the sphere down from out of the flow. What you are modelling is not physically possible and that is why the results are rubbish.

[sajeesh]

hi,

I am a beginner in fluent.i want to simulate the sphere moving in a fluid..the sphere is moving by the drag force and pressure force by the fluid initially.and then the motion of sphere modify the flow field ..finally the sphere reaches steady movement ...which method i should follow ...in ansys...plas help me ..or refer any tutorial to start with..somebody is telling VOF method.or particle transport method ..which i should follow pls help me.

[ghorrocks]

Try the fluent forum.

[sajeesh]

sir is this fluent forum

[ghorrocks]

The line
Home > Forums > CFX

Says CFX to me. The fluent forum is here: http://www.cfd-online.com/Forums/fluent/

[sajeesh]

thank you sir

[gowtham_donga]

Dear friends,
i am also working with the transient simulation of the flow past a cylinder but when i tried to plot the time vs force , i am getting a sudden jump at the end of the time .......i don't know why it is.........pls help me

[ghorrocks]

How are we meant to be able to work out what is going on when that is all you show? Can you post an image of what you are modelling, and the output file.

[vishnuambali@gmail.com]

I have tried a benchmark problem , transient flow past a circular cylinder with Re=100..I am getting the value of drag coefficient and frequency of shedding in the prescribed range but lift coefficient always falls out of the specified range...Even though my mesh is very fine near the cylinder...

What might be the problem ?? Pls help..

[ghorrocks]

FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F