rotational springs to connect link elements

  • 43 Views
  • Last Post 15 June 2018
julsmo posted this 07 June 2018

Hi all

I relatively new to ansys and have trouble creating rotational springs. the explanations in ansys help didn't help much.

I attached my ansys code and the iges geometry to import. I am trying to create rotational springs at each intersection. Was aiming to do that with combin39 elements but get weird movements as the nodes can't be coincident for that and I'm not sure if I implemented the springs right anyway.

Goal would be to create stiffness with the springs to get some kind of s-shaped curves when I pull on the net in the y direction.

appreciate any help. thanks!

 

code:

/prep7

 

et,1,link8

r,1,10,0.001

mptemp,1,0

mpdata,ex,1,,100000

mpdata,prxy,1,,0.2

 

esize,,10

lmesh,all

 

/solu

nsel,,loc,x,0

d,all,uz

d,all,ux

 

nsel,,node,,41

nsel,a,node,,1

d,all,uz,0

d,all,uy,20

 

nsel,,node,,100

nsel,a,node,,138

d,all,uz

d,all,uy

 

nsel,,node,,2

nsel,a,node,,80

d,all,uz

d,all,ux,5

 

nlgeom,on

alls,all

solve

Attached Files

Order By: Standard | Newest | Votes
peteroznewman posted this 08 June 2018

I recommend you use beams instead of links and rotational springs, since a beam is just an infinite number of links and springs.

If there is a reason to use links and rotational springs rather than beams, you can explain it further.

Attached is the Workbench ANSYS 18.2 version of your geometry from the IGES file above.

I don't use APDL much, but I did manage to see you wanted a 20 mm deformation in the y direction.

Attached Files

julsmo posted this 11 June 2018

Hi thanks for answering!

There is indeed a reason I use links. It's for my thesis and there is a previous thesis where all elements were modelled as beams but that apparently didn't picture the behavior of the cable net correctly. So I want to simplify it further assuming in the simplest version a cable is just a tnsion only link and add torsinal springs to get stiffness perpenicular to the cable since it's only unidirctional cables...

julsmo posted this 15 June 2018

Hi peter, just in case you see this. I am still working on the problem.. I have found a link element that works an dconnect it with spring elements, but can't couple the remaining dofs in a fixed way so both spring points move along the same distance. tried with a nonlinear mpc (rigid link).

attached is my skript and if you tell me how to attach an igs with my geometry, i can also do that

/prep7

 

nx = 2 !Maschenanzahl horizontal                                                     !!!ANPASSEN!!!

ny = 2 !Maschenanzahl vertikal

maschx = 40 !Maschenbreite                                                     !!!ANPASSEN!!!

maschy = 2 !Maschenhöhe

vorspannx = 5

vorspanny = 8

 

et,1,link8

r,1,10,0.001

 

et,2,combin14

keyopt,2,1,1 ! nichtlineare Berechnung

keyopt,2,2,6 !Feder rotz

r,2,10e10,,10e10

 

et,3,mpc184

 

 

mptemp,1,0

mpdata,ex,1,,100000

mpdata,prxy,1,,0.2

 

ksel,all

nkpt,,all

cm,feder1,node

*get,min1,node,0,num,min

*get,max1,node,0,num,max

 

nsel,,loc,x,0

nsel,a,loc,x,maschx*nx

nsel,a,loc,y,0

nsel,a,loc,y,ny*2

d,all,uz

nsel,,loc,x,0

d,all,ux

nsel,,loc,x,maschx*nx

d,all,ux,vorspannx

nsel,,loc,y,0

d,all,uy

nsel,,loc,y,ny*2

d,all,uy,vorspanny

 

 

 

ksel,all

nsel,none

nkpt,,all

cm,feder2,node

nmodif,all,,,1

d,all,rotz

 

*get,min2,node,0,num,min

*get,max2,node,0,num,max

 

 

type,1

real,1

esize,,10

lmesh,all

 

cmsel,,feder1

nsel,inve

nummrg,node

 

cmsel,,feder2

nsel,inve

nummrg,node

 

type,2

real,2

nsel,all

*do,i,min1,max1

*set,j,max1+i

e,i,j

*enddo

 

type,3

*do,i,min1,max1

*set,j,max1+i

e,i,j

*enddo

 

nsel,,loc,x,maschx*nx*0.5

nsel,r,loc,z,0

f,all,fz,1

 

/solu

nlgeom,on

alls

solve

 

peteroznewman posted this 15 June 2018

Unfortunately, I am a beginner with APDL.

I am highly skilled in Workbench and use an occasional Command object.

Someone else is going to have to help you with this.

Regards,
Peter

julsmo posted this 15 June 2018

thanks anyways!

Close