- مبلغ: ۸۶,۰۰۰ تومان
- مبلغ: ۹۱,۰۰۰ تومان
Many real-life applications of the Discrete Element Method (DEM) require a particle description which accounts for irregular and arbitrary shapes. In this work, a novel method is presented for calculating contact force interactions between polyhedral particles. A contact between two polyhedra is decomposed as a set of contacts between individual polygonal facets. For each polygon–polygon contact, an individual contact force is obtained by integrating a linear pressure over the area of its intersection. Both convex as well as partially concave polyhedra can be accurately represented. The proposed algorithm is validated by comparing to previously published experimental and computational gravitational particle depositions of identical cubes. Finally, the model is demonstrated in simulations of gravitational packing of various other polyhedral shapes.
4. Conclusion and outlook
In this study, a novel method was presented for simulating arbitrarily-shaped particles consisting of polygonal facets in the Discrete Element Method. Two bodies in contact are simulated as a set of interacting polygon-shaped contact primitives. As these primitives only need to contain local information about the geometry and mechanical properties, the method provides a very flexible framework to simulate contact interactions between particles of any shape and potentially non-uniform mechanical properties. Since there is no need for determining a unique contact point and normal unit vector for the contact between two arbitrary shapes, the method is not restricted to convex bodies and does not require disassembling arbitrary shapes into sets of convex bodies. It was shown that the computational cost scales quasi linearly with the number of contact primitives/particles and that – although introducing a clear additional overhead for “simple” shapes – the relative computational efficiency scales favorably when the particle shape becomes more complex. Furthermore, because each polygon–polygon contact can be individually resolved without information of the surrounding primitives, the method lends itself very well for parallelization.