Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Pre stress components.
Posted Mar 12, 2010, 3:02 p.m. EST 5 Replies
Please login with a confirmed email address before reporting spam
Greeting,
I have a brittle sample that will undergo thermal expansion, but before this I would like to give each element an initial stress state, I would like to give each element a randomly sampled value from a distribution (Weibull distribution). I was hoping to be told the best way of pre stressing the sample, from within the Comsol gui. I have the function written in the functions section, I am not sure how to apply it.
Also I was hoping someone could tell me the best way to adding element deletion when an element reaches a particular stress state the only way i could think of doing it is the have the young s modulus go to zero in the formula defining the modulus.
Thank you in advance.
Rizgar
Please login with a confirmed email address before reporting spam
I do not think you can give an individual element an initial stress, but you can define one for any sub-domain, or you would need to set an initial stress value as a function of rand(x,y) (to be checked), take a look at the "init" tab of the BCs.
You can also do an analysis and save the final state as inital condition for a restart/re-run
To change the stiffness, I beleive you should do as you said, set E to something small (perhaps not true 0)with a bolean condition, take a look at the examples for optimising 2D shapes, there are some good sample cases in there
Good luck
Ivar
Please login with a confirmed email address before reporting spam
As a sanity check if i generate a data file full of randomly sampled values then insert them as a data file function (somefn(x,y)) then for initial stress I supply that function. Would that put something like my random state onto the sample or will the interpolation and other steps like that turn it to garbage.
Please login with a confirmed email address before reporting spam
Please login with a confirmed email address before reporting spam
shure? did you do an "solve get initial value" ?
because if I put a displacement initial value of the type v=0.001[1/m]*x^2 for a simple 2D canteliever I get a initial quadratic deformation, so why not for stress ? (I cannot check as I have no ccess to COMSOL just now ;)
Good luck
Ivar
Please login with a confirmed email address before reporting spam
I think its because my data is very random when comsol tries to interpolate from one random value to the other it creates points that are not from the sampled distribution and makes it more noisy, while the function you are using is quite smooth.