CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   partially filled tank under acceleration (

Jean-Gabriel June 8, 2009 20:40

partially filled tank under acceleration
Hi, I'm not used to ansys CFX or Fluent but can someone help me with the next question.
As part of formula SAE team from ETS motorsport, we had an issue with the fuel tank in high speed cornering (1.5 to 1.8 g's) has a consequence the engine shut.

So, I would like to do CFD analysis to optimize the amount of fuel needed to ensure engine runnng under g's.

"How can i simulate a solid tank partially filled with fuel and air(under pressure 45PSI) in a acceleration state"

I don't need any animation only steady state result.

Thanks a lot as a new member of this forum I will post again for further information on CFX.

Rui June 9, 2009 04:07

That's a hydrostatic problem, isn't it?

stumpy June 9, 2009 06:57

I would just use a rotating reference frame - so you're simulating the car going around a circle. The rotation rate and the distance of the rotation axis from the tank will determine the g-force applied. Since both air and fuel are present it will be a free surface problem - see the "flow over bump" tutorial for how to set this up. make sure you use "Air ideal gas" as the material for the air phase, since the air will compress at 45 psi. Use a Reference Pressure of 45 psi. Use expressions to set the initial conditions for air/fuel volume fraction.

Rui June 9, 2009 07:20

Still . . .

Originally Posted by Rui (Post 218644)
That's a hydrostatic problem, isn't it?

ghorrocks June 9, 2009 08:07


Yes Rui, if all he wants is the steady state result then I can't see why you need CFD at all - just draw a plane which is perpendicular to the acceleration vector enclosing the volume of fuel. No CFD needed. See, I just saved you a few weeks worth of simulation.

But on the other hand there is NO WAY that a steady state surface will represent anything like what fuel in a tank in a race car wizzing around at 1.5g's is going to do. Here's a little experiment - fill a glass with water, give it a little push from the side and what happens? The surface rocks back and forth, maybe it spills over the top, maybe the flow forms breaking waves. You are kidding yourself if you think you can get anything meaningful with a steady state simulation.

If you are looking for a challenging CFD topic then do a transient CFD analysis of fuel sloshing in a tank. Unless you are an experienced CFD person with free surface modelling experience I guarantee it will take you several months before you get anything close to reality.

If you are looking to help your FSAE team design a better fuel tank than forget CFD and go to the library and look for some fuel tank design textbooks. You won't need CFD to come up with a better design.

Glenn Horrocks

Jean-Gabriel June 9, 2009 08:55

Thanks guys, the only reason why I would like to use the CFD simulation was because of the complexe geometry of the fuel tank. But, I don't have a month to spend in CFD simulation and so i will try theoritical calculation over a representative simpler geometry.

thanks again for the help

ghorrocks June 9, 2009 18:29


Complex geometry in the fuel tank? Sounds easy for a solid modelling package.

Glenn Horrocks

All times are GMT -4. The time now is 14:55.