Dear all,
I am working to replicate the Lifting Cylinder calibration test as in the paper by Roessler and Katterfeld (2019) called " DEM parameter calibration of cohesive bulk materials using a simple angle of repose test". As I am incorporating the Cylinder and the Ground in the STL files so that I can pull up the cylinder in a specific velocity, I noticed that with STL files, the memory requirement becomes significant larger. When I increased my CED values in the SJKR cohesion model to > 200K, my computer crashed after a long time of inactivity. As I want to simulate a highly cohesive material, I need to increase the CED between 400K-600K. I am not so sure about what went wrong in my simulation so that I am unable to increase CED values further, to compare my results with the results in the aforementioned paper.
1)Does anyone know why using the STL files takes so much of memory requirements? How can I reduce this requirement? I used gmsh to create my STL files?
2) Does anyone have an experience of simulating a highly cohesive materials using the SJKR model? As it seems that SJKR is based on the simple linear relationship, so for highly cohesive materials, the system can become unstable.
Thanks so much for your time. Really appreciate it if you can give me any advice.
Thao