Make blending wieghts in LUST user selectable
The proposal
LUST is a scheme that linearly blends linear interpolation with a linear upwind scheme using a 0.25 weight for for the latter. It is claimed that this
optimises the balance between accuracy and stability on a range of LES cases with a range of mesh quality.I don't necessarily disagree with this, but 0.25 upwinding is quite a lot in some situations, and insufficient in other. There seems to be no particular reason to not allow the user to choose the weight. I suggest that we do just that, with 0.75 remaining the default value when no user input is provided.
Beneficiaries
Mainly folk doing LES on unsructured grids.
Funding
I may be wrong, but implementing this should be pretty trivial and require minimal resources.
Bonus
The paper in which LUST appears to be introduced, different weights are tested. The paper is written by H. Weller. HILARY Weller :).
Weller, H. (2012). Controlling the computational modes of the arbitrarily structured C grid. Monthly Weather Review, 140(10), 3220–3234. https://doi.org/10.1175/MWR-D-11-00221.1
This reference should probably be added to LUST.H unless someone is aware of prior art.
Kind regards,
Timofey