High Order Approximation of Advection-Diffusion Problems on Polygonal
Meshes
March 26, 2025
3:00 pm PGH 646
Abstract
We consider solving second order advection-diffusion PDEs on general
computational meshes of polygons. We develop a general framework that
includes both high order finite element and finite volume
approximation techniques. Standard finite elements on polygonal
elements are either not available or lose accuracy. We develop new
finite elements that maintain accuracy on polygons by including
undistorted polynomials defined directly on the element. They possess
the minimum number of degrees of freedom subject to the accuracy and
Sobolev space conformity. The direct serendipity finite elements
approximate scalar functions (such as pressures, concentrations, and
saturations), while the direct mixed finite elements approximate
vector functions (such as velocities). Unfortunately, solutions to
advective problems can develop shocks or steep fronts, and thereby
lose Sobolev space conformity. We discuss the challenges of using
finite volume weighted essentially non oscillatory (WENO) techniques
on polygonal meshes. We also present a robust and efficient procedure
for producing accurate stencil polynomial approximations. We develop a
new multilevel WENO reconstruction with adaptive order that combines
stencil polynomials. The nonlinear weighting biases the reconstruction
away from both inaccurate oscillatory polynomials of high degree
(i.e., those crossing a shock or steep front) and smooth polynomials
of low degree, thereby selecting the smooth polynomial(s) of maximal
degree of approximation. Extension of the framework to three
dimensions is also discussed. Applications are given to tracer flow
and Richards equation.
2:00-2:30pm: talk for graduate students
Webmaster University of Houston
---
Last modified: April 11 2016 - 18:14:43