https://linen.dev logo
Then for some reason I stopped the simulation at 1...
# incompressible
l
Then for some reason I stopped the simulation at 1000 timesteps and restarted it, and pouf, the magU problems are back. The restart made something wrong
k
can you draw the component of U that's aligned with the pipe
this way you could see if there's a recirculation of some kind at the inlet
(happened to me already)
l
both at 1000 timesteps, seriously 😄
k
I suspect you use SST, low-Re mesh and 2nd order schemes for turbulence
l
k-eps, wall functions and first order lol
not but it really is the mesh
o
How the geometry looks?
l
I ran it before very very similar, geometry changed just a little little
k
well this can't be the problem, eh? 😄
l
look at this
I had that for the case before which just had a few degrees less of non ortho
the rest was strictly the same
typical pump with your 7 domains
no problemo now, surfaceLIC at inlet pipe. I had messy stuff with the hexa mesh though
anyway ye I stop complaining and I'll go fix my mesh
k
I don't think 76 degree should be that bad
l
The results are making sense
It just doest converge, p is stuck at 0.03 0.04, magU too big and wavy head. But it's still reasonable
k
all my simulations behave like that (except velocity at inlet), a pump just isn't a steady-state thingy
l
Yes I know but I somehow got such a nice convergence for my first polyhedral mesh that I thought it's the new norm with those
o
the poly you made it using cfmesh, polyDual or other mesher?
l
ANSA at work
fixed the mesh, but still have some troubles. I'm just gonna live with it 🙂