When studying the effect of confinement on flow past an object, some authors validate first their CFD model with experiment results for an unconfined flow then extend their numerical results for the confined case.
Is this approach valid only for this case or can be done for other types of boundary condition modification. For example validate the results of a flow past a fixed object and then extend the numerical results for the translating or rotating case.
Hi Hazel,
in my opinion you can not just modify the boundary conditions and expect the model to be still validated with the old experimental results. You can do something like this, if you have the so-called invariant, but in your case, you even change the experimental and so the numerical setup.
P.S.: Maybe there is some method to extend the model by some empirical approach, but it will surely have no physical meaning.
Best regards,
Antoni Artinov
Changing BC.s in the mathematical problem means you are prescribing conditions that are related to a different physical phoenomenon, for example changing from periodical to inflow/outflow or setting no-slip or slip condition at a wall. At the same time, once a physical condition is prescribed by a certain mathematical BC, you have to translate this in some numerical condition for your method. Sometimes, there is no unique implementation for doing that.
In conclusion, one have to assess both the physical BC and the numerical BC in controlled test-cases.
Thank you all for your answers. Let me please shape my question differently: In case there is no experimental results to validate the "modified boundary condition" case, is it an acceptable approach to validate the initial model with available experimental results and assume that the numerical results for the modified model are possibly good and could be discussed and published?
- Badges
- Science topic