[Jim-gineer]

Has anyone ever heard of CFD software which is also capable of modal analysis?

I am currently solving a problem regarding a free hanging water pipe in an underground storage cavern. The pipe is over 1000 feet in length and is suspended at one end. We believe that the pipe, which is actually many pipes of 40 feet in length threaded together, is reaching its resonant frequency. It is vibrating under normal operating conditions, due to fluid motion, causing the threads to jostle enough to let vapors from the underground cavern into the pipe, which is undesirable.

I have been successful in determining reasonable natural frequencies from ANSYS Workbench 11, but am looking for a way to analyze the effect of the fluid motion with regard to the frequency of the system. Ideally, running a modal analysis on a system with fluid motion involved would work, but I think that may be a bit ahead of where most CFD software is currently at.

Any advice regarding software, or a more practical approach to solving this problem is greatly appreciated.

[Jim-gineer]

To clarify: is there any way that CFD can be utilized to determine whether turbulent fluid flow is causing resonance? I guess that sums up what I am trying to find.

[acampbell]

The last couple of times we have looked at flow induced vibrations, we have come to the conclusion that the time it takes to get enough simulation time to generate the forcing functions required to perform a subsequent analysis is far too long (of the order of months for a small geometry). We even have had some undergraduate students look into it as part of a project.

Generally we would look at changing the supports etc to adjust the natural frequency of the structure.

This is all assumeing that the flow isnt doing something strange - like cavitating etc.

[Jim-gineer]

Thank you very much Mr. Campbell. I am actually an undergrad student myself, so I won't waste my time trying to solve my problem this way due to the complex nature of the geometry.

[meb]

I Jim,
I did similar studies in the past and I'm exploring rigth now a method that can help you.
We have developed an FSI module embedded inside Fluent. The dynamic behaviour is driven by the UDF (via a weak coupling and a mess of tricks to have a stable solution).
The success of simulation is strongly application related. We have simulated with success:
- 2d: motion of a flag, reed valve, elastic dam
- 3d: reed valve, a probe in a pipe (very similar to your application), a F1 front wing.
- 3d explicit: motion of a paper inside a printing device

We have also used exotic mixed approach using 2d or 3d beams that displace in the CFD solution their boundary.

I'm currently investigating the use of my new software (http://www.cfd-online.com/Wiki/RBF_Morph ) for the modal approach you are proposing. In fact the smoother is very fast and it can be used for modal superposition. You have only to pay the extra time required by Fluent for moving mesh solution (rezoning of the solution at each time step increase by a factor 2 the calculation time).

Regards.