[JuliaCan]

I want to simulate the flow in a hydro power plant. Before simulating a water hammer i want to make a steady simulation.

The problem is that my simulation does not converge. I think the reason is the poor mesh. When running the simulation the mesh statistic is quite bad. but i dont know how to improve it.
I tried to make the mesh more fine specially in the regions where losses will occure. I wanted to make a hex mesh instead of a tri but since i have to consider inflation layers it doesnt work.

+--------------------------------------------------------------------+
| Mesh Statistics |
+--------------------------------------------------------------------+
| Domain Name | Orthog. Angle | Exp. Factor | Aspect Ratio |
+----------------------+---------------+--------------+--------------+
| | Minimum [deg] | Maximum | Maximum |
+----------------------+---------------+--------------+--------------+
| steady | 18.2 ! | 357 ! | 747 OK |
+----------------------+---------------+--------------+--------------+
| | %! %ok %OK | %! %ok %OK | %! %ok %OK |
+----------------------+---------------+--------------+--------------+
| steady | <1 6 94 | 1 11 88 | 0 0 100 |
+----------------------+---------------+--------------+--------------+

Domain Name : steady

Total Number of Nodes = 161952

Total Number of Elements = 430410
Total Number of Tetrahedrons = 225410
Total Number of Prisms = 205000

Total Number of Faces = 70274


Does anyone have an idea how i can solve this problem ?

Thanks in advance

[JuliaCan]

... the mesh

[Steffen595]

maybe you can make several bodies and give them each a different mesh?
But I guess its more some boundary conditions, when it does not converge.

[ghorrocks]

Geometries like this (big regions connected to long thin pipes) need to be meshed very carefully. I recommend you use a multibody part so you can do sweep meshes in the long pipe regions, with high aspect ratio elements to reduce the element count. On the large regions you can use normal meshing technqiues.

[JuliaCan]

Hi,

Thanks a lot for your answer.
But how can I make an inflation when i use the sweep-method ?

[honcho62]

You could try using MultiZone and then inflating that method, that works very well for structured meshes with inflation.