Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2021-01-07 10:10:06

will.logie
Member
From: Canberra
Registered: 2017-09-23
Posts: 16

Spatial convergence study

Dear All,

I am looking at spatial convergence rates of code_aster for a pure eigenstrain (thermal stress) problem. The manufactured displacement field (NOEU_DEPL_R) I am comparing to integrates analytical thermoelastic stress/strain equations for the Bree (axisymmetrically heated) cylinder. The convergence rate for first order (QUAD4) mesh is as expected (gradient of 2), but for second order (QUAD8) elements it is very high (~6) until machine accuracy is reached.

Can anyone give me an idea as to why second order convergence in code_aster is not as FE convergence theory would suggest (gradient ~3), but in fact much better?

Kind regards,
Will.

Last edited by will.logie (2021-01-07 10:26:01)


Attachments:
mms_spatial.png, Size: 28.92 KiB, Downloads: 28

Solar Thermal Group
Research School of Engineering
Australian National University
http://stg.anu.edu.au/

Offline