[gravis]

Dear all,

I am running a simulation of pure airflow through a large cylindrical domain, about 5 m in diameter and 40 m in length.

I would like to compare the velocity field of this model to a scaled down CFD and physical model (1:50) with water instead of air, having the same Reynolds number.

A grid independence check and a Richarson extrapolation has been carried out for the scale CFD model, resulting in x number of nodes required for a certain level of error.

A independence study for the full scale model did not seem to be within the asymptotical range for the grids concerned.

My question is: would it be possible to in some way predict the order of the number of nodes that would be required for the full scale model, based on the grid independence check of the scaled down model?

This confuses me and also I am not sure about what sizes of turbulent scales to expect when comparing scale model with full-scale model.


I would be grateful to any suggestions,

Best regards,
Reine