A New Post-Processing Technique for Finite Element Methods with $L^2$-Superconvergence

Authors

  • Wei Pi School of Mathematics, Sichuan University, No. 24 South Section One, Yihuan Road, Chengdu 610065, China.
  • Hao Wang School of Mathematics, Sichuan University, No. 24 South Section One, Yihuan Road, Chengdu 610065, China.
  • Xiaoping Xie School of Mathematics, Sichuan University, No. 24 South Section One, Yihuan Road, Chengdu 610065, China.

DOI:

https://doi.org/10.4208/eajam.170119.200519

Keywords:

Finite element method, post-processing, least-square fitting, $L^2$-superconvergence.

Abstract

A simple post-processing technique for finite element methods with $L$2-superconvergence is proposed. It provides more accurate approximations for solutions of two- and three-dimensional systems of partial differential equations. Approximate solutions can be constructed locally by using finite element approximations $u$$h$ provided that $u$$h$ is superconvergent for a locally defined projection $\widetilde{P}$$h$$u$. The construction is based on the least-squares fitting algorithm and local $L$2-projections. Error estimates are derived and numerical examples illustrate the effectiveness of this approach for finite element methods.

Published

2020-05-04

Abstract View

  • 41450

Pdf View

  • 4572

Issue

Section

Articles