[Sparow89]

Hello everyone,

First post!

I am trying to replicate the experimental results of a paper ("The effect of flow oscillations on cavity drag" by Gharib and Roshko) using fluent. The paper uses an axisymmetric experimental model to examine the effects cavity geometry has on the frequency and drag of cavity flow oscillations. I am using this as a validation case before I move on to a similar problem for my own research.

I have tried two different geometries. The first is a 2D slice through the axisymmetric model. Imagine an ogive upstream followed by two identical cavities downstream that oppose each other separated by a sting (see attached image). The second model is trying to make use of the 2D axisymmetric function to reduce cell count and flow problems.

I am using the DES with SST k-w solver, Reynolds number should be around 24,000, y+ is 1 or less, wall spacing around boundaries and in/around the cavity is uniformly set to 6e-5 m. Mesh contains around 9M elements for the first geometry and half for the second. The top and bottom boundaries are slip walls and I am using a velocity-inlet pressure-outlet. Residuals are set to 1e-6. Solver settings are set to second order where applicable and default everywhere else. I am using surface monitors to measure dynamic pressure and drag.

What I am finding is that, for the first geometry(as pictured), there is a large pressure difference between the two cavities when I would expect them to be close to identical. Also, the flow oscillations are sporadic and not consistent with experimental results at all.

For the second geometry the cavity flow is pretty much the same as for the first case. Very sporadic flow and no consistent oscillations.

I had one case that yielded good results, but it only seemed to enter a good flow regime after a restart.

I would appreciate any advice or thoughts that may help!

Please let me know if there is any additional information that would be helpful.
Thanks,
Stephen

[IMG][/IMG]