DPDsim

An implementation of several variants of dissipative particle dynamics, a coarse-grained molecular dynamics method to simulate mesoscale structures.

Available methods:

1. Standard dissipative particle dynamics (DPD), source: Groot and Warren, JCP, 1997
2. Many-body dissipative particle dynamics (MDPD), source: Warren, PRE, 2003
3. A generalised many-body DPD with freedom in defining local density and wrapping function (GMDPD) presented in Vanya and Elliott, PRE, 2020

This implementation of DPD/MDPD allows non-bonded particles only. For more complex simulations of polymers, we recommend DL_MESO or LAMMPS.

Code versions

Project is mainly written in Python. Key bottlenecks involving force and energy computation are optimised as follows:

• Enhancement by Numba library (experimental)
• A Fortran module linked via f2py (about 10x faster)

Post-processing tools to compute density profiles and radial distribution functions are in utils.py.

(NB: There are also ad hoc versions in pure Fortran and Julia not developed anymore.)

Usage

Standard DPD with Fortran-optimised code:

from dpdsim import DPDSim

sim = DPDSim(implementation="fortran", steps=1000, thermo=100)
sim.create_particle_inputs(kind="pure")  # alternative is 'binary'
sim.run()

Other force fields (MDPD, EMDPD, GMDPD) are handled in the same way, eg:

from dpdsim import GMDPDSim

sim = GMDPDSim(implementation="fortran", steps=1000, thermo=100)
...

Access to key physics-computing functions:

sim.compute_pe()             # returns potential energy
sim.compute_ke()             # returns kinetic energy
sim.compute_local_density()  # computes MDPD local density stored in "sim.rho2"
sim.compute_force()          # computes force, virial and the stress tensor diagonal

Post-processing from within Python works as follows:

from dpdsim.utils import compute_profile, compute_rdf

# run with DPDSim object "sim"
r_pr, pr = compute_profile(sim, 0, 1, N_bins=50) # 0 for x-coord, 1 for particle type
r_rdf, rdf = compute_rdf(sim, 1, N_bins=50)      # 1 for particle type

Troubleshooting

In case of questions, bugs or ideas for improvement, please contact me at peter.vanya~gmail.