PADI

Title of research project:

Description of mechanical particle interactions by an implicit algorithm

Reference: F1R-ING-PUL-08PADI

Project Leader: Bernhard Peters

Research Team: Kasra Samiei (Ph.D.), Stefan Maas, Andreas Frommer

Research Domaine: Engineering

Project period:

Partners : (Bergische Universität Wuppertal)

Objectives:

Demands for the extension of computer simulation methods to granular systems of increasing size and complexity cannot be met solely by developing better hardware and consuming more computer time; needs for conceptual improvement of the existing algorithms are becoming more and more obvious. An implicit algorithm, in principle, offers an improvement to existing methods of the Molecular Dynamics simulation of granular materials. In the proposed methodology, the inter-particle forces are to be described as functions of positions and velocities of the granular particles in contact at the current time step. This will form a system of differential equations implicit in positions of the particles. The numerical solution to the system will give the positions and accordingly other useful information (velocities, accelerations) to predict the motion of the granular materials. The implicit algorithm is to be implemented in C++ programming language and integrated into DPM (Discrete Particle Method) software, an advanced simulation tool derived from DEM (Discrete Element Method) to simulate the motion of granular matter. The method is to be tested in some common devices involving granular matter motion namely hoppers, forward acting grates, reverse acting grates and rotary kilns. The results will be verified against currently more common methods of simulation of granular material i.e. the so called explicit approach and in certain cases against the available experimental results.

Results:
A literature survey conducted suggests that the experience with implicit methods in the context of molecular dynamics has been limited and reporting mixed, i.e. both positive and negative, results. Some works discuss the possibilities of reducing the implementation costs and developing efficient algorithms.