MercuryDPM–SPH: An Open-Source Polydisperse SPH Framework with Viscoplastic Rheology

  • Singh, Chandrabhan (University of Twente)
  • Askes, Harm (Maastricht University)
  • Weinhart, Thomas (University of Twente)

Please login to view abstract download link

MercuryDPM–SPH is a newly developed open-source Smoothed Particle Hydrodynamics (SPH) framework embedded within the discrete particle code MercuryDPM. The solver is designed for polydisperse particle systems, complex rheology, and future SPH–DEM coupling, targeting engineering applications such as fresh concrete, slurries, and particulate suspensions. By reusing MercuryDPM’s hierarchical grid neighbour search, the method achieves near-linear computational complexity even for highly poly-disperse discretisations. The framework employs a weakly compressible SPH formulation with multiple kernel functions, density update schemes, and equations of state. Density can be computed using summation, the continuity equation, or a hybrid cyclic approach with optional Shepard renormalisation, while pressure is evaluated using linearised or nonlinear Tait equations. Numerical stability is ensured through Monaghan artificial viscosity and an optional δ-SPH diffusion term. The solver supports Bingham and Herschel–Bulkley rheology with Bercovier–Engelman and Papanastasiou regularisations. In addition to the conventional continuum stress formulation, a DEM-inspired pairwise shear force model is introduced, where viscous and viscoplastic interactions are computed directly between particle pairs using tangential relative-velocity projections. This provides a thermodynamically consistent dissipative mechanism, improves robustness for yield-dominated flows, and forms a natural basis for SPH–DEM coupling. The solver is validated using rheometer, hydrostatic, and free-surface tests, showing quantitative agreement with analytical results. Additional Couette and Poiseuille flow benchmarks are currently under development. The code is released under a BSD-3-Clause licence with full source and reproducible cases.