fracture mechanics, j-integral

[francal]

I'm trying to solve cases of fracture mechanics known in reading with ansys,
Would someone know where to find a cint command guide, since in ansys it is not exhaustive at all? ? ?

Thank you.

 

[sciolingelao]

Bye francal
I took a look at the various pages of ansys v.12 online help.
Now I don't know... I have even found examples of apdl listings that, I think, explain in a quite simple way the use of the " cint" command.
assuming that you did the same thing, I invite you to explain yourself better, what exactly are you looking for?
Hi.

 

[francal]

I'm sorry, I didn't explain so well.
the cint command by itself is easy to use, I wanted a detailed guide to understand how the percentage error varied when changing the parameters.
For example, the questions I'm trying to answer are:
1- even with the j-integral mesh needs a quarter point as well as for the kcal? theoreticalmete the answer should be no, since the j-integral is an energy criterion and is not based as the kcalc on the ctod (positions of the apex)
2-What changes to the change of contour in the cint command? theoretically nothing since the j-integral is an integral independent of the domain...

if someone could give an answer I would be grateful:smile:

 

[sciolingelao]

francal your request has "dipped me " not little.
between the pages of my old university texts concerning fracture mechanics I found a formula that, contrary to what you suppose to point 1, league j to ctod:
j=m * sigma ( yeld stress) * ctod
m takes values in a range from 1.15 to 2.95

 

[francal]

for:
1) kcalc uses the movements of 3 particular points to calculate the intensity factor of ki voltages
2) Integral j as energy release rate
received showed that the energy release rate for non-linear elastic materials containing a
Slot can be written in the form of line integral independent from the path.then the line integral is equal to the energy release rate for a nonlinear elastic material
under almost-static conditions.
so the integral j can be seen both as an energy parameter and as a parameter
of the intensity of the tension state.

My question is very simple:
since the j-integral does not look at agi shifts around the apex of the clique (but only to those of the contour),Why would I need a quarter point in the j-integral?Thank you.

 

[sciolingelao]

Hello, francal.
I only answer you now because only now ends my working day (started at 5.00 ).
We try to reason a bit, it will be difficult since I am not very lucid. . .
go to point 12.2 of the ansys online guide.
go to the j-itegral formula ( 12-1 )
You know anything? sigma and tau appear
Now, as you used to point the reference axes (see Fig. 12.7 ) what would happen for x or y tending to zero, namely: if our path coincided with the apex of the clique what would we get?
I believe, at this point, that the reason why we still model the **** tip and ***** front with the help of the singular elements, so much with elements that have nodes located at their quarter points is to overcome the situation of singularity that we would find near the apex of the clique.
I don't know. I can't say he's getting drunk.
What do you think?

 

[francal]

Now, as you used to point the reference axes (see Fig. 12.7 ) what would happen for x or y tending to zero, namely: if our path coincided with the apex of the clique what would we get?
I believe, at this point, that the reason why we still model the **** tip and ***** front with the help of the singular elements, so much with elements that have nodes located at their quarter points is to overcome the situation of singularity that we would find near the apex of the clique.

We're all over this.
in fact if set:

cint,contains,1

they select the contour closer to the apex, and I must say however the values are high

if instead set

of which,6

I take a contour away from the apex, and the results have a massim error of 3%! ! !

If you decide to do a test with a mesh totally random, then without quarter point, with a contour away from the apex...overwhelmed by the chair because the Russians are quite good.

I'd like to be sure, but with the cint command you don't need quarter point...

appearance confirmed

 

[sciolingelao]

Good morning francal
I found something that seems to confirm your argument, that is, the need not to use "singual elements ".
I would like to point out that the material I am recommending to you is accessible to all (no pirated ):www.google.it insert ansys cint.
the first link is a power point file " ansys linear elastic fracture mechanics example "
consult slide n° 8:demonstration problem: ansys j-integral method
I take the opportunity to wish you and all our friends my best wishes for a holy Easter.