TY - JOUR
T1 - Stabilized finite element methods for steady advection-diffusion with production
AU - Harari, Isaac
AU - Hughes, Thomas J.R.
N1 - Funding Information:
The authors wish to thank Ken Jansen for helpful discussions on turbulence modeling and Greg Hulbert for valuable comments on allowable errors in time-integration schemes. In the course of this work, Isaac Harari was supported in part by the Center for Absorption in Science, Ministry of Immigrant Absorption, State of Israel.
PY - 1994
Y1 - 1994
N2 - Finite element methods for solving equations of steady advection-diffusion with production are constructed, employing a linear source to model production. The Galerkin method requires relatively fine meshes to retain an acceptable degree of accuracy for wide ranges of values of the physical coefficients. Numerous criteria for improved accuracy via Galerkin/least-squares are proposed and examined in the entire parameter space. Of these, several provide the lowest error in nodal amplification factors, each in a certain range of ratios of physical coefficients. Guidelines for required mesh resolutions are presented for Galerkin and Galerkin/least-squares, showing that in large parts of the parameter space substantial savings may be obtained by employing Galerkin/least-squares. Galerkin/least-squares/gradient least-squares, a new variation of the Galerkin method, is designed to provide exact nodal amplification factors, offering accurate solutions at any resolution. Numerical tests validate these conclusions.
AB - Finite element methods for solving equations of steady advection-diffusion with production are constructed, employing a linear source to model production. The Galerkin method requires relatively fine meshes to retain an acceptable degree of accuracy for wide ranges of values of the physical coefficients. Numerous criteria for improved accuracy via Galerkin/least-squares are proposed and examined in the entire parameter space. Of these, several provide the lowest error in nodal amplification factors, each in a certain range of ratios of physical coefficients. Guidelines for required mesh resolutions are presented for Galerkin and Galerkin/least-squares, showing that in large parts of the parameter space substantial savings may be obtained by employing Galerkin/least-squares. Galerkin/least-squares/gradient least-squares, a new variation of the Galerkin method, is designed to provide exact nodal amplification factors, offering accurate solutions at any resolution. Numerical tests validate these conclusions.
UR - http://www.scopus.com/inward/record.url?scp=0028431239&partnerID=8YFLogxK
U2 - 10.1016/0045-7825(94)90193-7
DO - 10.1016/0045-7825(94)90193-7
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:0028431239
SN - 0045-7825
VL - 115
SP - 165
EP - 191
JO - Computer Methods in Applied Mechanics and Engineering
JF - Computer Methods in Applied Mechanics and Engineering
IS - 1-2 C
ER -