Viscosity Calculations with Hybrid Particle-Field Molecular Dynamics Simulations

This thesis presents a systematic investigation of field-based simulation methods for modeling simple fluids, focusing on the Hybrid Particle-Field (HPF) approach and its combination with Multi-Particle Collision Dynamics (MPCD). Using the Lennard-Jones fluid as a benchmark, conventional molecular dynamics (MD) with explicit pairwise interactions is compared to HPF and HPF-MPCD models, evaluating their ability to reproduce static and dynamic fluid properties. It is shown that HPF accelerates simulations and reduces their run time by up to two orders of magnitude. This speedup is achieved by mediating interactions through a coarse-grained density field. However, this efficiency comes at the cost of altered structural correlations and artificially fast single-particle dynamics. Additionally, incorporating MPCD into HPF restores momentum-conserving collisions, enabling accurate control over hydrodynamic properties such as the shear viscosity. It is demonstrated that key transport coefficients can be tuned to match Lennard-Jones reference values by adjusting MPCD parameters, and a symbolic regression model is provided, relating viscosity to collision frequency, temperature, and density across a wide range of state points. Further, the critical role of numerical implementation is highlighted: a momentum-conserving gradient discretization (8-point central difference) is essential for the physical fidelity of HPF-based simulations. Performance benchmarks reveal that HPF and HPF-MPCD offer substantial speed advantages over conventional MD, especially for large systems or long-time simulations, with only modest additional costs for MPCD. Taken together, these results establish HPF-MPCD as a flexible and efficient framework for simulating fluids with tunable dynamical properties, bridging the gap between fast, field-based models and fully atomistic simulations. The methodology outlined here provides a robust foundation for future studies of complex fluids and soft matter systems, with applications ranging from high-throughput parameter scans to multiscale modeling.