SBIR/STTR Award attributes
As device designs become more complex and incorporate new, advanced technologies and materials, accurate simulations of multi-physics phenomenology become critical to ensuring that such design and analysis efforts are successful. The diversity of problems of interest to DOE and wider audiences includes examples such as electric motor/generator design with reduced rare-earth content, bioelectromagnetic (bio-EM) processes for advanced imaging and future EM-based therapies, plasma-wall interaction and damage models, and superconducting EM systems for plasma confinement. Developing advanced computational physics solvers, particular multi-physics and large-scale electromagnetic software, is critical to solving these challenges in an accurate and timely manner. These advanced capabilities must also be augmented with modern high-performance computing (HPC) enabled software, all while improving user accessibility. Current multi-physics, EM, and HPC software like DOE’s MFEM require substantial expertise across multiple disciplines to be used effectively. MFEM, part of DOE-funded CEED in ECP and SciDAC/FASTMath, is a modular finite element (FE) programming framework developed for a unique audience of expert-level scientists and engineers with programming background. In order to expand the accessibility of this great DOE resource, this project focuses on incorporating MFEM into three existing software products, enabling scientists and engineers without computational programming expertise the opportunity to leverage MFEM in a user-friendly capacity. Overall, this combination of technologies will make HPC- enabled simulations more accessible to non-expert users and democratize EM and multi-physics simulation capability all while increasing addressable problem size through HPC use. Future expansion may yield complete cloud-based HPC simulation through integration with cloud CAD platforms. This work has two major objectives. First, the team will demonstrate the feasibility of integrating MFEM into a multi- physics software package that currently lacks some of the key benefits provided by MFEM, namely high-order accuracy and HPC solutions. Second, the team will demonstrate integration with a computational EM software package to improve efficiency of multi-scale solutions for problems like bio-EM and RF signal integrity, an important problem for 5G RF-microelectronics. These two integration efforts will provide confidence toward the ultimate goal of a commercial HPC-enabled simulation software with built-in access to expert-level finite element solution modules. The team will demonstrate both efforts on a set of test problems to validate the approach. Achieving a successful demonstration will provide confidence that the abstraction approach developed on this program is broadly applicable to multiple software products and pave the way for integration with other software products. The long-term goal, initiated by the work of this Phase I program, is to provide users with multiple software products built upon a powerful HPC-enabled finite element modeling capability. The team envisions at least three near-term products, built upon existing software. After Phase II, the team will improve the services offered to customers through a design-as-a-service model, for which the team will leverage the software abstractions, HPC capability, and robust optimization to provide automated device design services.
