Skip to content

Latest commit

 

History

History

Creep

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Creep: Bar under constant load

Tested with CGX 2.19 / CCX 2.19

Model based on the official ccx example beamcr.inp but with modified material parameters and load application

  • Large displacements
  • Visco-elastic material
  • load application via *coupling with *distributing
File Contents
pre.fbd CGX input for preprocessing
solve.inp CCX input (constant force)
solve2.inp CCX input (constant true stress)
peeq.plt Gnuplot script for the creep curve (strain)
disp.plt Gnuplot script for the creep curve (displacement)
test.py Python script for the complete analysis

Setup

Parameter Value Meaning
width 1 cross section dimension in mm
length 8 length in mm
le 1 node distance

The parameters are defined in pre.fbd. The mesh and the sets are generated using

> cgx -b pre.fbd

The model consists of a bar of cross section 1 x 1 mm and 8 mm length (longitudinal direction Z). It is fixed at z = 0 in z-direction while allowing for deformation in x and y directions.

There are two versions of load application:

  1. solve.inp as a constant concentrated force to the control node of surface Szl (creating constant total force)
  2. solve2.inp as a constant negative pressure to surface Szl (creating constant true stress)

Fully integrated 20-node elements are used, because reduced integration would lead to hourglassing with just a single element over the cross section.

The material parameters are taken from literature and represent magnesium alloy AZ91 at 200°C.

The simulation consists of a *visco step of 6.7 hr (constant total force) or 20 hr (constant true stress) duration.

Solving

Constant total force

> ccx solve

The simulation doesn't reach the designated total time because of excessive creep rate when the cross section becomes too small.

> monitor.py solve

Constant true stress

> ccx solve2

The solver history can be documented with

> monitor.py solve2

Postprocessing

The equivalent creep strain is written to the .dat file.

This is converted to a tabular text file suitable for gnuplot.

> dat2txt.py solve
> dat2txt.py solve2
> gnuplot peeq.plt

Equivalent plastic strain and deformed shape of the profile: