Hi everyone,
I am experimenting with a microclimate (Wind) and Urban Heat Island (UHI) study for the South African Astronomical Observatory (SAAO) in Cape Town using a custom geospatial framework. I am currently on a Community License, which means I need to be extremely efficient with my core-hour usage and mesh sizes.
The Challenge:
My current Incompressible mesh (I haven’t even gotten to the heat transfer yet) is hitting 16M cells even at low fineness, largely due to how the mesher treats the TIN terrain facets and 0.25m building intrusions. I am frequently seeing “Face-face inconsistency” and “Min Edge Length : 0” errors.
Because I need to eventually run thermal comfort and UHI indices, I can’t completely ignore the surface data, but 16M cells is not feasible for a Community account.
Questions for the Power Users:
-
Given the core-hour limits of the Community License, what is the best “target cell count” for an urban area of approximately 500m x 500m? I am currently aiming for 4-7M, but the terrain keeps driving it higher.
-
Is there a way to coarsen the terrain mesh specifically while keeping the building-air interfaces sharp?
-
Would simplifying my TIN terrain (reducing triangle count in my CAD before export) resolve the “Face-face inconsistency” more effectively than adjusting the Small Feature Suppression in SimScale?
Current Settings:
-
Fineness: 3
-
Small Feature Suppression: 0.001m
-
Gap refinement: 0.5
-
Global Gradation Rate: 1.22
Any tips on getting a “lean and mean” mesh that stays within Community limits while providing enough data for thermal analysis would be huge.
I’ve run a simulation and failed and will continue experimenting. The work is here