The investigator studies the accuracy and reliability of large eddy simulation (LES) of turbulent flows. The research on LES includes modeling, analysis, algorithm development, numerical analysis, and computational testing. The recent mathematical development of approximate de-convolution models for turbulence shows this approach to be an excellent path for developing accurate, stable and predictive LES models. The research on finding better LES models is thus based on the approximate de-convolution framework. The research on matching algorithms to models uses an improved understanding of models and their numerical analysis to develop better algorithms. The investigators research studies reliability of simulations by developing analytical foundations of models and algorithms directly from the Navier Stokes equations and by computational estimation of model sensitivity. Training Ph.D. students properly to contribute to this area is a large part of the proposed effort.
The accurate and efficient simulation of and extraction of reliable information from turbulent flows is a central computational challenge in many important applications including global change estimation, biomedical device design, energy efficiency improvement, optimization of industrial processes, safety estimation of the new generation of nuclear reactors, pollution dispersal and abatement and security issues involving dispersal of biological and chemical agents. All the above applications require accurate and reliable numerical simulations of complex flows and thus progress in these applications needs efficient and accurate models of turbulence and fast and reliable numerical methods. This project involves the development and testing of models for turbulent flow which can address essential difficulties in the modeling and simulation of complex flows. This research studies the physical and mathematical foundation of the models as well as efficient computational methods for performing numerical simulations of the models. The aim of the research is to advance the efficiency, reliability and accuracy of the simulations of turbulence which form a core difficulty in these, and many other, applications.
