Efficient Initialization of Low-Density Polymer Systems Using Dissipative Particle Dynamics
Faculty Mentor Information
Dr. Eric Jankowski, Boise State University
Presentation Date
7-16-2026
Abstract
Simulations give unique insight into polymer structure and dynamics, but are difficult to initialize. To simplify and expedite system initialization we use dissipative particle dynamics. While initializing high-density systems is particularly challenging, we explore initialization of low-density polymer systems to determine whether initialization protocols developed for higher densities are transferable. We generate initial configurations over a range of system sizes, quantifying structure with radial distribution functions and bond lengths. We find simulation parameters representing repulsion magnitude, viscosity, and bond stiffness have the largest impact on overall run time, but are generally slower than for more dense systems. We explore the implications of relaxing the criteria for interparticle separations, which enables faster initialization, but increases the chances of unstable configurations when used in production polymer simulations. We investigate method parameterizations for systems up to millions of particles and observe linear time dependence on system size. We compare graphics processing units against CPU architectures and find 60,000 particles to be a practical limit for initializing systems in under 15 minutes. These findings demonstrate the broad transferability of our initialization techniques across a broad range of densities and provide a foundation for further work in polymer science.
Efficient Initialization of Low-Density Polymer Systems Using Dissipative Particle Dynamics
Simulations give unique insight into polymer structure and dynamics, but are difficult to initialize. To simplify and expedite system initialization we use dissipative particle dynamics. While initializing high-density systems is particularly challenging, we explore initialization of low-density polymer systems to determine whether initialization protocols developed for higher densities are transferable. We generate initial configurations over a range of system sizes, quantifying structure with radial distribution functions and bond lengths. We find simulation parameters representing repulsion magnitude, viscosity, and bond stiffness have the largest impact on overall run time, but are generally slower than for more dense systems. We explore the implications of relaxing the criteria for interparticle separations, which enables faster initialization, but increases the chances of unstable configurations when used in production polymer simulations. We investigate method parameterizations for systems up to millions of particles and observe linear time dependence on system size. We compare graphics processing units against CPU architectures and find 60,000 particles to be a practical limit for initializing systems in under 15 minutes. These findings demonstrate the broad transferability of our initialization techniques across a broad range of densities and provide a foundation for further work in polymer science.