Multiphysics & Multiscale Fluid Dynamics

Some of the most challenging fluid flow problems involve the interaction of multiple kinds of physics occurring over disparate lengths and time scales, such as fluids with particulate impurities, combustion of multi-species biofuels, and electrostatically charged spray painting processes.

Identifying and appropriately coupling physical models to achieve accurate, stable, and computationally tractable results is a daunting task. At Berkeley Lab, we are developing state-of-the-art mathematics and numerical methods for a variety of multi-physics multi-scale fluid flow problems. Examples include coating flow and evaporation of multi-layer paint films, multi-scale models of foam dynamics, and coupling the macroscale gas dynamics of bubble rearrangement to the microscale dynamics of the soapy liquid inside the thin films. We are developing fluctuating hydrodynamics models that capture thermal fluctuations in various fluid types, including charged fluids and fluids containing complex molecules, such as polymers or proteins.

Tackling these multi-physics problems also requires significant development of advanced computing methods and simulation on some of the world’s most powerful supercomputers. So our researchers are also developing highly-scalable algorithms targeting exascale computing and techniques such as reduced order modeling to help explore the vast parameter landscape typical of applications in manufacturing and industry.

Projects

 In addition to the projects listed under the Astrophysics & Cosmology and Environmental sections,  projects focusing on multiscale fluid dynamics include:

MuMSS: Multiscale Modeling and Stochastic Systems

At macroscopic scales, fluid dynamics is governed by partial differential equations that characterize the behavior of the fluid in terms of smoothly evolving fields that represent density, momentum, and other characteristics of the fluid. However, at atomic scales, fluids are discrete systems composed of individual molecules whose dynamics are governed by complex interaction potentials. The discrepancy between these two descriptions is manifest at the mesoscale. Our researchers are developing models and algorithms that can efficiently and accurately model complex fluids at the mesoscale. Contact: Daniel Ladiges (Ladiges on the Web)

Optimization of Rotary Bell Atomization

Our researchers are working with PPG Industries – one of the world’s largest paint manufacturers – to couple advanced mathematics with high-performance computing (HPC) resources to model the paint drying process and guide the development of new energy-efficient coatings systems for the automotive industry. Contact: Robert Saye (Saye on the Web)

Machine Learning Accelerated Models

Our Machine Learning Development Team is using machine learning models within an AMReX framework to accelerate the time-to-solution of computational kernels without reducing accuracy. Contact: Andy Nonaka (Nonaka on the Web)

MFiX-Exa

MFiX-Exa is a multi-phase flow simulation tool for the exascale. Contacts: Weiqun Zhang, Andrew Myers

News