Simulation of Mechanical Process Engineering Applications with the DEM Software ROCKY using the GPU Technology
Speaker: Clovis Maliska (ESSS)
Recent and ongoing improvements of the GPU Compute Technology is enabling and expanding the usage of numerical simulation tools in many areas. A prominent example is the Discrete Element Method (DEM) which forms the basis of Computational Body Dynamics (CBD), where the dynamic behavior of many solid interacting particles or bodies is predicted. In practice, the method is being used to analyze and improve mechanical process engineering equipment, as for example granular material handling. The GPU technology, compared to CPU , allows to run simulations with a much larger number of particles in significant shorter time. And, alternatively, the increased GPU compute power enables the modelling of discrete elements by faceted geometries. This provides much more flexibility of DEM for the simulation of real shape bodies compared to the common practice where spheres or a collection of glued spheres are used.
The presentation covers two topics: First, it summarizes past and ongoing trends of the General-Purpose GPU technology (GPGPU) and how the benefits of the GPU technology have been implemented and used in the development of the DEM program ROCKY . This part also includes a short overview of some relevant and unique features of the code. Second, several examples are presented for illustrating the productive gain in running on GPU and in modelling real shape bodies. The examples cover mechanical process engineering applications, such as bulk material transport, sieving and breakage, pharmaceutical tablet treatment and quality assurance, manufacturing processes for complex production parts such as grinding, and others.
Finally, an outline is provided how ROCKY is coupled with FEM and CFD codes to address Multiphysics applications where the interaction with structural analysis and fluid dynamics is important.
Info: http://talks.cam.ac.uk/talk/index/120130
Contact: Dr James Dean (jd362@cam.ac.uk)
Registration: https://goo.gl/forms/UEFQ3Zie4is5VerG3 (Space is limited, registration is required.)