Coarse-Grained Mapping of Fluid Particles via Evolutionary Fuzzy Clustering: Membership-Evolution Term as a Pressure...

Area of Science:

• Computational physics
• Chemical physics
• Materials science

Background:

• Coarse-graining (CG) methods bridge atomistic and mesoscopic scales in fluid simulations.
• Fixed CG mappings struggle with the dynamic, unbundled nature of fluid particles.

Purpose of the Study:

• To develop an improved CG method addressing limitations of fixed mappings.
• To investigate the impact of evolving fuzzy particle-cluster memberships on CG accuracy.
• To establish a bottom-up approach for pressure correction in CG fluid dynamics.

Main Methods:

• An entropy-regularized fuzzy clustering algorithm with temporal smoothness constraints was employed.
• The force-matching framework decomposed cluster interactions into particle-interaction and membership-evolution terms.
• The method was applied to Lennard-Jones (L-J) fluid and water systems.

Main Results:

• An intermediate fuzziness level optimized structural features in radial distribution functions.
• The membership-evolution term consistently contributed a repulsive force across systems.
• Including the membership-evolution term in CG dynamics simulations restored system pressure for L-J fluid.

Conclusions:

• The proposed fuzzy clustering CG method enhances structural descriptions and pressure accuracy in fluids.
• The membership-evolution term offers a physically grounded pressure correction mechanism.
• This work provides insights for designing advanced coarse-graining strategies for fluid systems.