Department: Mechanical & Aerospace Engineering
Faculty Advisor(s): Sutanu Sarkar

Primary Student
Name: Daniel A Nelson
Email: d6nelson@ucsd.edu
Phone: 916-595-4613
Grad Year: 2015

Defining precise and accurate flow boundaries in complex turbulent flows has been a significant challenge in the analysis of mixing and muti-phase transport. Traditional Eulerian analysis has not provided a method with the necessary precision to identify the topologies of coherent structures which give rise to the transport characteristics in a given flow. On the other hand, recent Lagrangian methods have lead to the development of theories that do indeed provide a mathematical criteria for flow boundaries. One such method is the theory of Lagrangian Coherent Structures (LCS). An LCS field is quantified by determining the finite time Lyapunov exponent (FTLE), which is a measure of the maximum stretching in the fluid. However, low order FTLE determination coupled with low order fluid simulations leads to an inaccurate identification of coherent structures. This in turn has consequences on the accuracy of the mixing and transport characteristics. This work utilizes a high-order spectral element method to simulate turbulent flows via direct numerical simulation (DNS), wherein the smallest flow scales are fully resolved. Work also includes the development of a new algorithm to determine FTLEs utilizing spectral element methods as opposed to traditional finite-difference schemes to improve accuracy of FTLE computation. Recent results verify the spectral accuracy of the new algorithm and also point to the possibility for more efficient FTLE computation by requiring fewer particles to trace in order to achieve desired accuracy.

Related Links:

  1. http://attila.sdsu.edu/~jacobs/
  2. http://nelson-dsc.com

« Back to Posters or Search Results