slow after mesh for enhanced wall treat DPM

joshkemp

I ran my original case using realizable k-epsilon with non-equilibrium wall function with about 900k cells. Then I used the DPM model to add spray to the domain and the discrete phase iterations took a few more hours to run .

Then I realized my wall y+ values were all too low and it was going to be too hard to try to get the cells large enough to be above Y+ of 30, so I decided to switch to enhanced wall treatment with all y+ values less than 1. The resulting mesh was again around 900k cells. I ran the new mesh again and had no problems getting convergence on the airflow. Then I turned on the DPM model again and tried to iterate for the spray. I let it run about 24 hrs and it still had not finished one DPM iteration.

I am confused as to why it the DPM iteration is taking so long. Like I said, the mesh is about the same size and both cases use exactly the same DPM setup. I am reading the same initial distribution in from file. Both cases were ran on the same 20 node computing cluster. The airflow on the new mesh did not take any longer to converge than the original mesh. All of the cells are large enough so that the largest droplet will not occupy more than 10% of the volume.

Any Ideas?