Solving solid mechanics problems with material plasticity

General discussion about Elmer
kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 25 Nov 2019, 16:44

Elmer
Elmer.JPG
(29.12 KiB) Not downloaded yet
NASTRAN (matches calculated solution)
nastran.JPG
(85.43 KiB) Not downloaded yet

mika
Posts: 107
Joined: 15 Sep 2009, 07:44

Re: Solving solid mechanics problems with material plasticity

Post by mika » 26 Nov 2019, 19:04

I checked that the StressSolver and ElasticitySolver give consistent results for small loads. Could there be some difference in the specification of the load for the different software? What is the resultant couple vector you expect to obtain at the fixed end?

-- Mika

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 26 Nov 2019, 20:33

The solution expected is below. The moment at the fixed end is end force times length of beam. I will re-check the total end load.
solution.PNG
solution.PNG (42.69 KiB) Viewed 397 times
Kevin

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 26 Nov 2019, 20:52

The error was mine. The end load in elmer the way I defined it is force/length. Putting an end load in as 3E6 is only 3E5 do to the height and line being 0.1 long. Putting the end load in at 3E7, for a total force of (3e7 * 0.1) = 3E6, results in good answer.
result.PNG
(52.05 KiB) Not downloaded yet
Again thanks.

Also i am getting close to posting my first plastic material UMAT model. 2D elements bi-linear stress strain curve. Preliminary testing looks good, I am refining it, then planning a validation run. I am outputting invariant values to a text file, since Elmer is not yet calculating them for UMAT.

Kevin

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 30 Nov 2019, 13:58

A bi-linear 2D plastic material model UMAT has been developed and tested. UMATLib_Plastic.F90
bilinear_model.png
bilinear_model.png (10.79 KiB) Viewed 381 times
It is located here with several test cases. Further testing and development will be ongoing.

https://github.com/mrkearden/UMATLIB


Planning to include compression as well as tension. Planning 3D solid elements.

Since ElasticSolver does not support calculating invariant values such as mises, the UMAT routing does and outputs to text files.
I plan on passing these back to ElasticSolver with state variables and perhaps use SaveData to output.

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 02 Dec 2019, 13:02

Add another subroutine to the UMATLib_PLastic.F90 bi_linear_sym assumes the compression stress-strain curve is the same as the tensile stress strain curve. bi_linear has yielding in tension only. test case also add to github.
I also added 9 state variables, elastic strain,plastic strain,total strain,elas max prin, plas max prin, total, elas mises, plas mises, total
symmetric yield
sym_plasticity.png
(43.14 KiB) Not downloaded yet
tensile yield only
plastic_stress.png
(41 KiB) Not downloaded yet

raback
Site Admin
Posts: 3512
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Solving solid mechanics problems with material plasticity

Post by raback » 04 Dec 2019, 00:47

Hi,

Nice work! We are on the process of streamlining the ElasticitySolver a bit and this should make new material models to have even greater impact.

-Peter

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 06 Dec 2019, 14:33

I have greatly simplified the bi-linear UMAT. It now checks the mises stress to determine yielding. Model convergence is greatly improved.

https://github.com/mrkearden/UMATLIB

see the github for file and the test case called coupon
result.png
result.png (22.68 KiB) Viewed 327 times

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 07 Dec 2019, 01:38

This approach made it simple to adapt to 3D or 2D elements, yield occurs in either compression or tension, since it is based on mises stress.
UMATmises_yield.F90
(8.57 KiB) Downloaded 19 times
case.sif
(2.43 KiB) Downloaded 22 times
strain.png
(42.73 KiB) Not downloaded yet

kevinarden
Posts: 392
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Solving solid mechanics problems with material plasticity

Post by kevinarden » 09 Dec 2019, 13:29

I have updated this again to improve file writing and calculations of principal and mises. This is done in a solver call PlasticSolver. PlasticSolver is ElasticSolver with the addition of calculations of principal and mises and output to files called output_stress.txt and output_strain.txt. It only does it for one specified integration point on one specified element.

The UMATmises_yield.F90 works with ElasticSolver, but only mormal stress and strain will be output.
UMATmises_yield.F90 with PlasticSolver will output the above.

https://github.com/mrkearden/UMATLIB

PlasticSolver is built by
elmerf90 PlasticSolve.F90 -o PlasticSolve
and callrd in the sif by
Procedure = "PlasticSolve" "ElasticSolver"

Post Reply