hard points

[volaff]

Bye.

Could someone tell me what hardkeypoiuts are? ?

if I have a text file with hardkeypoints how do I change their location? ? ?

practically in the model were shaped rivets with beam elements4, rivets scattered between two knots of ends.

In fact I should change the location of such nodes and, therefore, rivets, but I have the text file used to generate hardkeypoints.

Thank you all!

 

[stefano.garbin]

Given that I never used them, first of all their name is hard points.
find them in the mechanial apdl guide - modelling and meshing guide - solid modelling - paragraph 5.2.2
there is a good explanation. basically associate these lines or areas and the mesh that is then generated will always put us on a knot.

 

[volaff]

Thank you very much for the answer.

one last thing (for now...xd): they gave me a model already "ready" of an airboard consisting of two overlapping panels and riveted along the common side.

rivets were modeled as beam4 elements and an element was created between the extreme knots of the rivet through a type text file:

finish
/prep7

*I,i,1.8
ksel,s,,12032+i
ksel,a,,5319+i
♪
*get,n1,node,nxth
*get,n2,node,nxtl
type,4
e,n1,n2


*end


I have the need to change the position of the knots and, therefore, rivets.
I wanted to know politely if it is better to do it in graphic maneira or always through the help of a text file (although I did not really understand how to do it).

Thank you very much
Have a good day!

 

[stefano.garbin]

Now I'll explain what the file you posted to me does:

finish
/prep7
*do,i,1,8
ksel,s,,,12032+i
ksel,a,,,5319+i
nslk
*get,n1,node,n,nxth
*get,n2,node,n,nxtl
type,4
e,n1,n2
*enddo

with the first two lines enter the preprocessor environment
then part a do cycle repeated for 8 times.
in the do cycle are selected two keypoints (ksel), then select the nodes associated with them (nslk and succ.) and create a beam4 element between those two knots.
Note that the selection of keypoints takes place by number and not by xyz location in this file.

I think at the end of the fair you might want to do everything graphically.

 

[volaff]

Thank you very much for the suggestion.
In fact I should eliminate the nodes of extremity + the hedge between the knots and do it again for every single rivet.... it will take me two months I will know...dedication! !

thank you very much for the availability!

ps: you already explained the file posted what he did and I was very helpful..maledetta thesis di Laurea! !

 

[stefano.garbin]

But how many rivets do you have to do?
if they are so many then it is game force find an automated way to manage everything and not via graphical interface.

 

[volaff]

Well the rivets are quite a bit here because a procedure has been used by text files.
Unfortunately some rivets are "storted" so some knots are remade.. .

 

[stefano.garbin]

then you have to clean the mesh, recreate with text files all hardpoints and then regenerate the mesh.

 

[volaff]

or my god, I'll kill myself then! !

cmq to what I understood the hardpoints were used to get tensions and deformations inside the panel so, in fact, I would not just devote to changing the ends of the rivets to solve the repoblem?

Or didn't I get a bat?

thank you very much for the availability and courteous! !

 

[stefano.garbin]

why not take it as an opportunity to learn something new?

 

[volaff]

definitely! !

the first help will be ansys' help surely.
next week after the exam I have to do, I'll take it easy and patient.

So let me clean up the mesh, crave the hardpoints again and then remake the mesh?
I always accept the council of an expert!

 

[stefano.garbin]

If you want or you can try to post some screenshots of the problem you're sinking so we understand each other better.

 

[volaff]

That's great. ! !
Should I post the model image with its rivets then?

 

[volaff]

while I create screenshots and similars, you could tell me where I can look for something on text files to use with ansys apart from the software's help! ?

Thank you so much!

 

[stefano.garbin]

Carol volaff, I'll give you some advice I think is important.
try to make from scratch a model in classic ansys, with a little preprocessing (geometria, materials, loads etc.), of solution and postprocessing, all from graphical interface.
then from menu click on list->files->log file
you will come out big text file where all the things you did, all the commands you gave, all the zoom, rotate, pan etc. operations were recorded.
Clean up this file from all the graphical operations and you will no longer have what ansys does. for each operation there is a command (the language is called apdl - ansys parametric design language) whose arguments are completely parameterizable.
then on the guide studied the whole chapter entitled "ansys parametric design language guide".
will open a new world: difficult but extremely powerful.

 

[volaff]

Thank you very much for the answer.
the problem is this: I wanted to remake the model already internally but the prof told me not to take initiatives and to change the model that he gave me.
I said this as soon as I support the exam I have to do shortly I throw myself on that chapter you reported trying to pull out something positive.
thank you very much always for the courtesy and availability!

ps: the model is beautiful and ready to make but the rivets must be straightened.
I don't know if the logifle of the model that I have reported all the operations related to the axes used by those who created the model or reports the operations of the axes installed on my pc.

I'll let you know!

 

[volaff]

Hello everyone, I'm back here with more doubts than before.
I press that I could not in any way get the log file of who made the model in question.

I contacted my prof who told me not to change the position of all hardpoints (which should only be used to detect deformations) but only those related to beam4 elements (reverse).

between the material I was given there are some test files used to create both hardpoints and rivets.

going to plot the only beam4 elements I have the situation reported in figure where it is evidently seen that some rivets are "storted".

the prof suggested me to work with text files as it is more comfortable seen the high number of rivets.

At this point I wondered how it was possible to go to select the rivets ( beam4 element between two knots) incriminated as if you plotted the number of screen nodes would not understand anything.

If, instead, you want to select medainte commands (apdl) the incriminated nodes trying to straighten them as I can know which extreme (node) should be straightened?

the panel is defined by section and the different tread files I have (which concern the various sections) are all of this type:

finish
/prep7

*I,i,1,40

Ksel,s,,11499+i
ksel,a,11699+i
♪
*get,n1,node,nxth
*get,n2,node,nxtl

type,4
e,n1,n2

*end

by itself I understand what the script of the text file does, the problem is to select the right nodes.

Sorry I've been long.

I attach some picture of the panel

 

[stefano.garbin]

hi volaff, from this situation you can get out with some programming in apdl.
I would do something like this:
1-select all rivets (esel,s,ename,,beam4)
2-for each rivet (with a *do cycle) select the ends knots
3-then select the hardpoints attached to those knots
4-recording x,y,z coordinates of hardpoints couple
5-if the rivet should be aligned to the z axis check if the x,y coordinate pairs are the same
6-if so I do not go to change x,y so that they are the same
7-end cycle *do, regenerated mesh and relaunch solution

clear the strategy? Does that sound correct? Let me know.

 

[volaff]

Thank you very much for the answer

for command 1 is fine:

ksel,s,type,,4, (type 4 elements are beam 4 i.e. rivets)

the problem is that I have 2377 beam4 elements.

Would the procedure also be revised for rivets or with a single script?
I did not understand the numbering used by the model creator.

at the end the first step is to straighten the rivets of that central area where the two panels overlap (the problem is right there).

the real doubt that I axle is another: I do not believe that at the ends of the rivets are attached hardpoints!! !

in the sense that the prof said that we used hardpoints only to go to detect deformations where there are no knots, so in fact the hardpoints I have to use them only to detect deformations and should not return to the nodes to change! ! !

You're always kind.
Maybe tomorrow I try to work with one rivet and see what I can pull out!

Good night and thank you again!

 

[volaff]

then the text file on the rivets that "attack" the two panels in the central area is as follows:

finish
/prep7

*do,i,1,40 start cycle do with counter "i" from 1 to 40 (*)

ksel,s,,11499+i!select keypont number 11499+i
ksel,a,,11699+i also select the keypont number 11699+i
nslk!select the knots attached to these keypoints
*get,n1,node,n,nxth!memorizzo in n1 the number of keypoints greater
*get,n2,node,n,nxtl!memorizzo in n2 the number of keypont minor

type,4!setto as current element type 4 (beam)
e,n1,n2!creo an element between n1 and n2


*end
!(*) I carry out 40 times the sequence of operations between *do and *enddo varying "i"
(that's exactly what I had reported yesterday...xd)

in this way the text file goes to create 40 rivets.

small question: how can I see only these 40 rivets anizchè all beam4 elements?

so I don't get confused with the rivets?

if you used:

esel,s,type,,4, should then select all beam4 and not only those in question..... I looked for the chapter in question but I didn't understand a great deal! !

See you soon!