Reply To: How to create 2D triangle ？
w.r.t. resolution: The N constant (which as you say is commonly between 10 – 35) is only a useful proxy for the actual resolution that results of the spatial discretization delta and the domain size. This delta is computed by dividing some reference length (chosen to fit what is modeled by each simulation case) divided by the N constant.
w.r.t. inclined geometry / pipe: If my understanding of what you want to simulate is correct, you would be better served to keep the geometry axis aligned s.t. it is nicely discretized by the regular grid and model the inclination by rotating the external force field. This also should get rid of the cuboid decomposition problems (which point to some issue in domain setup).