First of all: I read the topic “Optimize Simulation” and tried to transfer the suggestions to my mesh, but in my Opinion the main Problem there ist the overall quantity of cells. I don´t think that this counts for my mesh with something like 650K cells.
I finally managed to run my multiphase Simulation and with the help of a colleague we could speed it up by the factor of two.
We did this with a new moderate hex dominant mesh instead of the standard algorithm. Unfortunately it´l still take ~3,5 days/s of simulation (In my case it will be stopped after 30000s anyway).
The solver log now says that the velocity exceeds 500 m/s which ist way too fast. Expectation ist a maximum of less than 0,5 m/s at the inlet. Furthermore the Volume Ratio is 105. A.f.a.i.k. this should be below 100. I solved this by meshing again with “fine”, but didin´t yet run a simulation with this.
Besides the meshing, what else can I do to speed up simulation? I have to say that the Model is relatively large (I think). It´s a manhole (grid chamber) with 6m in diameter and comparable depth. The inlet and outlet is 1,8m in diameter each. Inflow is 0,775 m³/s. May this actually be too large for the utilised 16 cores max?
The moderate mesh ist still running. After finishing, I will start again with the fine mesh. Maybe I can add more information after the first timesteps.
I would be grateful for your Ideas.