Graphics Processing Units based Computing System for Emerging Hypersonics Solvers

Abstract

High-speed propulsion is essential for realizing hypersonic flight. In this regard, airbreathing scramjet engines, and detonation-based devices are gaining interest for powering a wide range of vehicles. Both these propulsion concepts involve complex interaction of shocks, turbulence and chemical reactions in their flowpath. In order to realize an efficient, robust and reliable design, deep insight into the core physics and tools for design optimization are needed. However, the multiphysics multiscale nature of these flows pose considerable modeling challenges, especially due to the resource intensive nature of solving the governing equations. The focus of this project is to develop a graphics processing units (GPU) based computer cluster that helps develop a new generation of hypersonics solvers that can not only extract physics but help in online modeling and design optimization. GPUs provide high compute density and low power consumption, which will allow new strategies for solving the governing equations. In particular, GPUs are particularly useful when the physics problem is cast in terms of vectorized algorithms. For instance, artificial neural networks based algorithms are highly efficient on such systems. The GPU-based system will enable end-to-end workflows for concept-to-design optimization in an automated manner.

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 29, 2024
Source ID
FA95502310067

Entities

People

  • Venkatramanan Raman

Organizations

  • Air Force Office of Scientific Research
  • Board of Regents of the University of Michigan
  • United States Air Force

Tags

Fields of Study

  • Physics

Readers

  • Computational Fluid Dynamics (CFD)
  • Distributed Systems and Data Platform Development

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • Hypersonics