Different options, huge differences in results, which option's result is accurate?
The same xml file, the same computer, I use Chrono to calculate, it takes about a day; using the DEM option that comes with DualSPHysics, the calculation takes a fixed time step of 1e-6, and the calculation takes half an hour.
This is not terrible, the most terrible thing is that there is a huge gap between the DEM calculation results and the Chrono calculation results, it is only a difference of options, nothing else has changed.
When comparing the two calculation results with the literature, there are two extreme phenomena, neither of which can be compared with the literature results, one is much larger than the literature results, and the other is much smaller than the literature results.
why is that? Are there references to the relevant verification properties for DEMs in DualSPHysics, and which option yields more reliable results? Is the code not mature enough or is there something wrong with my operation?😨
Comments
Have you read about DEM and Chrono? Have you understood the equations solved for each one? Have you checked the details of the implementations? Have you read papers comparing DEM and CRHONO?
Shooting questions before reading and getting information is wasting a lot of time... yours and ours!!!
Solid interactions computed by DEM and by CHRONO are different in terms of detection algorithm so differences should appear.
DEM solves interactions in terms of particles, where usually dp is the minimun distance for collision between particles. This means that results is really dependent on the resolution. On the other hand sliding can not be properly solved since smooth surfaces are not discretized as smooth planes using particles in DEM.
On the other hand CHRONO will use collision detection based on geometries (.vtk, .stl, .obj, etc) and impacts take place for a minimum distance (defined in our code in terms of dp: distancedp) so that results are not dependent on the SPH resolution and sliding can be solved accurately.
More discussing can be found in:
Canelas RBC, Crespo AJC, Brito M, Domínguez JM, García-Feal O. 2018. Extending DualSPHysics with a Differential Variational Inequality: modeling fluid-mechanism interaction. Applied Ocean Research, 76: 88-97. doi:10.1016/j.apor.2018.04.015.
Regards
@Alex Ok, thank you for your reply, I have read the paper you provided before, maybe I don’t understand enough, I will read it carefully again, thank you very much for your reply and suggestion!