DIS (Dynamic Interactions Simulator)

Key Advantages over Competing Solutions
  • DIS modeling formulation uses an explicit time-integration solver that can maintain time-accuracy with negligible "drift" for long simulation times.
  • The explicit solver is “embarrassingly” parallel. The evaluation of the element flux and vector which is the step that takes more than 95% of solution time can be done in parallel (on multiple CPUs or GPUs) by simply dividing the finite elements between available processors. Other solution steps are done on a per-node-basis and can also be done in parallel.
  • Single code for CFD/SPH, DEM and FE simulations: DIS solves the coupled CFD, DEM and structural dynamics equations using the same code without a need for “gluing” two or more different codes together. In the “gluing” approach, the codes are run simultaneously with the outputs passed between the codes during each time step. This approach however suffers from a number of limitations:
    • Creating a combined model in different software systems unavoidably involves entering repetitive information. Thus more time is needed from engineers to create the model and more care is needed to ensure that the repetitive information is consistent between codes.
    • Setting the proper boundary conditions between the codes is usually not trivial since the codes expect the information that is provided to be compatible with their formulation.
    • For accurate solutions data should be exchanged between codes at the smallest time step level. Transferring this data wastes a lot of computer resources.
    • The solution produced is usually not very accurate mainly due to loss of information (energy and momentum) due to interpolation of the solution fields at the boundaries.
    • Post-processing of the results is also not trivial since each software code might output data in a different format.
  • Improved computational efficiency: This is due to the fact that the finite element, CFD, SPH, and DEM calculations are all done using the same code which avoids the inefficiencies associated with code “gluing” and back and forth exchange of data between the original software and the add-on module(s).
  • Single framework for pre-processing, simulation, and post-processing.