fcFEM - FEA from start to finish

About the development of the FEM module/workbench.

Moderator: bernd

User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: fcFEM - FEA from start to finish

Postby HarryvL » Thu Feb 07, 2019 8:39 pm

... and how about uniform pressure on a curved surface. E=1000MPa, p=1MPa, D=10mm. Theoretical and calculated increase in diameter 0.01mm


Sphere.jpg
Sphere.jpg (55.52 KiB) Viewed 441 times
User avatar
bernd
Posts: 8329
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: fcFEM - FEA from start to finish

Postby bernd » Thu Feb 07, 2019 10:26 pm

:shock:


PrzemoF wrote:
Thu Feb 07, 2019 3:47 pm
So we'll have an internal solver?! :o
HarryvL wrote:
Thu Feb 07, 2019 5:12 pm
Yup
:D
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: fcFEM - FEA from start to finish

Postby HarryvL » Fri Feb 08, 2019 3:21 am

HarryvL wrote:
Thu Feb 07, 2019 8:39 pm
... and how about uniform pressure on a curved surface. E=1000MPa, p=1MPa, D=10mm. Theoretical and calculated increase in diameter 0.01mm
To be precise: calculated change in diameter: 0.0100001941635mm, so an error of 0.001941635%. Not bad considering the 10-node terahedron has 6-node parabolic (and not spherical) faces.
User avatar
PrzemoF
Posts: 2649
Joined: Fri Jul 25, 2014 4:52 pm
Contact:

Re: fcFEM - FEA from start to finish

Postby PrzemoF » Fri Feb 08, 2019 7:44 am

That will be absolutely amazing! I'm currently involved in scripting compilation of CalculiX and seeing things like fortran '77, spooles/arpack code from the 90's hacked together with some shady perl scripts and shell macros is like digging in an old basement. There might be dead bodies ;) I know that CalculiX won't be replaced any time soon, but having internal solver would be so much better!
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: fcFEM - FEA from start to finish

Postby HarryvL » Fri Feb 08, 2019 7:57 am

... and much easier to extend with special features.
User avatar
bernd
Posts: 8329
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: fcFEM - FEA from start to finish

Postby bernd » Fri Feb 08, 2019 10:28 am

PrzemoF wrote:
Fri Feb 08, 2019 7:44 am
That will be absolutely amazing! I'm currently involved in scripting compilation of CalculiX and seeing things like fortran '77, spooles/arpack code from the 90's hacked together with some shady perl scripts and shell macros is like digging in an old basement. There might be dead bodies ;) I know that CalculiX won't be replaced any time soon, but having internal solver would be so much better!
+1 and On one side CalculiX has a good community and lot of people use it, but the development still is a one man show of guido (even after over 20 years of development (AFAIK)). This is not future save IMHO.
Jee-Bee
Posts: 1950
Joined: Tue Jun 16, 2015 10:32 am
Location: Netherlands

Re: fcFEM - FEA from start to finish

Postby Jee-Bee » Fri Feb 08, 2019 11:24 am

bernd wrote:
Fri Feb 08, 2019 10:28 am
+1 and On one side CalculiX has a good community and lot of people use it, but the development still is a one man show of guido (even after over 20 years of development (AFAIK)). This is not future save IMHO.
Why is it not open?
User avatar
PrzemoF
Posts: 2649
Joined: Fri Jul 25, 2014 4:52 pm
Contact:

Re: fcFEM - FEA from start to finish

Postby PrzemoF » Fri Feb 08, 2019 11:31 am

I think it is open, but writing code in fortran 77 is some serious mental exercise. A random code sample from CalculiX:

Code: Select all

if(nelem.eq.nelem1) then
!     
               ang1s=(1.41d0-0.00594*alpha1)*alpha1*pi/180
!     
               cang1s=dcos(ang1s)
!     
               if(alpha1.le.22.5) then
                  lam11=0.0712*alpha1**0.7041+0.37
                  lam12=0.0592*alpha1**0.7029+0.37
               else
                  lam11=1.d0
                  lam12=0.9d0
               endif
               zeta=lam11+(2.d0*lam12-lam11)*(V1V0*a0a1)**2
     &              -2d0*lam12*V1V0*a0a1*cang1s
               zeta=zeta*(W0W1)**2
!     
            elseif(nelem.eq.nelem2) then
For me it's just one step away from https://en.wikipedia.org/wiki/Brainfuck
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: fcFEM - FEA from start to finish

Postby HarryvL » Fri Feb 08, 2019 11:42 am

I am not planning to replicate Calculix. It’s just that I have many wishes on my todo list that would be impossible to implement without full access to and understanding of 3rd party code. I am now implementing buckling and stress stiffening (the opposite of bucking) and interface elements. This would take me months Calculix and days in something as simple as fcFEM.
Jee-Bee
Posts: 1950
Joined: Tue Jun 16, 2015 10:32 am
Location: Netherlands

Re: fcFEM - FEA from start to finish

Postby Jee-Bee » Fri Feb 08, 2019 1:17 pm

PrzemoF wrote:
Fri Feb 08, 2019 11:31 am
For me it's just one step away from https://en.wikipedia.org/wiki/Brainfuck
i think Brainfuck (i had never heard about it :o) is worse ;)
would be great if Calculix comes at github. would speedup the development i guess