project help:mesh2d

[Sturm]

Hello to all,it is the first time I use a forum,I hope to have written in the right section,I have a problem with a university project,I have to perform an analysis on a plate(I am using plane82 elements),I have to get a very precise result of the maximum equivalent voltage(the result is required in mpa and it is required precision up to the 3/4 decimal digit to quano seems...) and I can not get an acceptable convergence by manually honing mesh.
the only command I found through the help was adapt but gives me the following error "adaptive looping is stopped because error norms are zero. make sure that analysis and element types support error calculations".
Therefore I wanted to know if it is the command mentioned above to be used in these situations or if there is a simpler method for meshare at best my plate.
I want to point out that I have already tried to use both plane42 and 82,quadrilaterals and triangles, to divide my area into two parts in order to get mesh more accurate and thick in the critical area, to use a unique mesh with various control parameters on the dimensions of the elements.all these solutions give me identical results only until the first decimal digit but then begins the zig zag...
I hope someone can help me, thank you in advance!

 

[John]

How is the plate bound?
Do you have to determine only tensions or shifts?
Does the plate have a thickness?

would be useful a desccrition of geometry and analysis that you want to do... .

the adapt command I have never used it, even because it can only be used with purely static analyses if you already start doing something else does not go...
it is possible to make a *do cycle and to make mesh in a time when it is what the adapt command does with the difference that the program goes to fix the mesh where the energy error is greater, but this can also be done with an analysis of convergence of results

 

[Sturm]

Hello! Just to start thank you very much for the answer!
then the plate is bound according to y on the lower side and on the left side according to x, has a radius hole 1 mm at the top right and thickness of 2 mm. inside the hole you have a pressure applied on its right semicircumference.
I have to determine only tensions.
I hope to have explained sufficiently well!!

 

[John]

already better, although I did not understand how it is bound and without geometries, material properties etc etc. is a little problem to give results :biggrin:....
cmq were in you would use shell63 as elements and I would go to cool heavily near the hole and where there is a great variation in the efforts.. .
cmq the value of the sigma where you must determine it?

 

[Sturm]

I also looked on English forums and eventually found someone who had my same problem...that is, too thick knit does not give correct results for the equivalent sigma of von mises that tends to grow to increase the density of the mesh with ansys.The problem is that, in this way, finding the convergence of the solution becomes really difficult manually.
do you know by chance what elements the adapt command works best?
tomorrow I try to see if it changes anything with the shell63 elements you suggested.
Thank you again!

 

[John]

That's why I asked you where you went to "see" the variation of the sigma of von mises, because if there are problems of singularity the value will never come to convergence (it happens precisely that increasing the mesh increases the value continuously). .
try to understand where you want to analyze the value of the sigma, better if it is far from the constraints so there are no problems with the singularities and you should get to a convergence. . .


I put a quick-made file to figure out how the adapt command can work, basically choose a tolerance value for the structural error value something and change the mesh automatically... but I repeat it is much more convenient to use *do cycles so you decide where to stick, in the area you need
finish
_

/prep7

!
et,1,shell63! element shell63
mp,ex,1,210000!modulo young [N/mm^2]mp,prxy,1,0.3 !coeff poisson

!diversi spessori shell
r,1,1

blc4,0,0,10,10

lesize,all,20
amesh,all


lsel,s,loc,x,0
dl,all,all

sfa,all,pres,0.05

finish

/solu

antype,0

adapt,,7,,1

solve