Nonlinear Geometric Models: Algorithms, Analysis, and Computation

University Of Maryland, College Park, College Park MD

Investigators

Abstract

Fabrication and manipulation of new and smart materials, particularly in the strategic areas of nanotechnology and biotechnology, require understanding of nonlinear phenomena governed by geometric partial differential equations (PDEs). At small scales, say micro and nano scales, surface tension and bending effects dominate bulk effects, thereby making the actuation and control of small devices a reality. This leads to scientifically interesting and technologically useful configurations and dynamic behavior. Examples abound in biomedical sciences (drug delivery vesicles, cell encapsulation devices, and sensors) and engineering (photovoltaic devices, optics, energy storage, micromotors, microgrippers, microvalves, and adaptive deformable mirrors). However, microfabrication is time-consuming, expensive, and often erratic, which makes the development of predictive computational tools of paramount importance in engineering and science. This project deals with modeling, analysis, and computation of geometric problems of interest in materials science, biophysics, plasma physics, and robotics. It enhances modeling and prediction capabilities and helps educate students and postdocs in exciting, mathematically and computationally challenging, and practically relevant areas of contemporary research. Capturing the essential behavior of nonlinear phenomena with the simplest and crudest models is fundamental in science and engineering. This allows for understanding of basic mechanisms, the design and implementation of efficient numerical methods for simulation and control of devices, and the analysis of both models and algorithms. These crucial aspects of modern research are incorporated into the following four intertwined projects: geometric PDEs with constraints (bilayer actuators and prestrained films, shape optimization for plasma confinement, and fully nonlinear PDEs); actuation of complex fluids (liquid crystals actuated by electric fields and temperature, and ferrofluids actuated by magnetic fields); nonlocal models (efficient solvers for linear and nonlinear fractional diffusion and stochastic control); a posteriori error analysis and adaptivity (high-order methods, fractional PDEs, and free boundary problems). Numerical treatment of nonlinear geometric PDEs is a formidable scientific challenge due to the dynamic deformation of geometries, the presence of strong nonlinearities, and the development of self-penetrating structures and topological changes. Efficient algorithms should optimize and balance the computational effort and thus capture small scales without over-resolving others, thereby leading to accurate interface description. This project develops structure-preserving finite element methods (FEMs) with a posteriori error control (adaptive FEMs) and multilevel solvers, which allow for the resolution of problems with very disparate space-time scales with relatively modest computational resources. The roles of geometry, nonlinearity, nonlocality, and adaptive approximation permeate the research, from basic questions in numerical analysis of nonlinear PDEs to applications in strategic areas of national interest. Graduate and postdoctoral students participate in the research of the project. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.