Learning Forum

[javat33489]

Hello. I am solving a jet pump problem in Fluent. In the usual calculation, everything is considered excellent for 2000 iterations. There is a zone of negative pressure and I put the calculation with cavitation, but the calculation does not work:

What could be the problem?

I've tried both stationary and non-stationary methods. T

he WOF model is implicit.

Energy is on.

The carrier liquid is water (the first phase), the second phase is water vapor.

Ke model with refined walls.

I also tried SST, the result is the same.

Help please.

[Rob]

Cavitation can be difficult to predict well. Are you running transient, and how well resolved is the mesh relative to both the geometry and pressure field?

[javat33489]

When calculating cavitation, is it necessary to use a transient?

[Rob]

Not always, it depends on the level of cavitation and what it does to the flow. In your case it may come down to how the flow balances over the various channels from high to low pressure. 

[javat33489]

What recommendations can you give? Thank you

[javat33489]

What does it mean? Can I ignore negative pressure? But then how to interpret it? Or do you mean such moments can be considered in stationary mode? How to understand when it is time to switch to non-stationary mode?

[Rob]

If you ignore something without understanding you have a problem. If you decide something isn't significant from a position of knowledge then you need to justify that. I cannot comment on interpretation: that's for you and your supervisor to figure out. 

As I also don't know what the device is, it's hard to judge. To add, transient and steady refer to the solution of time. "Stationary" would translate as not moving: English isn't a simple language so I just want to clarify if parts of the solid/surface mesh are moving or not. 

In terms of switching from steady to transient, there is no simple answer. If the boundaries are transient, it's more straight forward. If the transients are flow induced it depends if they matter and if they alter the answer to the question you want to solve using CFD. I may run a model that exhibits transient behaviour as steady as I don't care about small changes in the flow: if the changes are large but don't alter what I am worried about I may also neglect them. However, I've also been doing this a while so have a fairly good understanding of how something should behave.