Aligning the geometry with the cell

[MahmoudZew]

Hello,

Do we really need to align geometry obstructions, such as walls, blocks, etc. with the mesh cells?

As I understand it, FDS resolves an obstruction only if it occupies at least half of the cell size. If the obstruction thickness is less than half a cell, FDS treats it as a thin obstruction: no heat transfer calculations are performed, although it still acts as a barrier to the flow.

Given this behavior, what are your thoughts on strict geometry-mesh alignment, especially in building environments where room dimensions often do not align neatly with mesh cell size multiples?

I’d be interested in your perspective on best practices in such cases.

[drjfloyd]

Heat transfer calculations are performed for all obstructions, thick or thin.

[MahmoudZew]

So what do you suggest? Should I align my obstruction with the mesh cells, or does it not matter?

[mcgratta]

FDS automatically snaps the obstructions to the grid. For this reason, some people like to enter the actual coordinates of the geometry and let FDS snap it Some, however, prefer to simplify the geometry before entering it in, choosing, for example, a 10 cm grid. With this second approach, you do not risk your geometry being snapped to the grid in an unexpected way.

[drjfloyd]

The major errors that are going to result from misalignment are free air volumes of compartments, surface area of compartments, and area of openings. If you are reasonably well resolved, the fact that a room snaps to an extra grid cell is not going to be a large source of error.