Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

Failed to compute elasto-plastic variables

Please login with a confirmed email address before reporting spam

Hello,

I'm trying to solve an Structural Mechanics problem in 3D:

In fact, it's a concrete block that will be loaded in compression. I have introduced a Drucker Prager model defined by the equation:

F = alpha*Tr_s+sigma_eqv-K;

where,

alpha = (fc-ft)/(fc+ft)
Tr_s = sx_smsld+sy_smsld+sz_smsld
sigma_eqv = sqrt(0.5*(sx^2+sy^2+sz^2)+sxy_smsld^2), et
k = 2*(fc*ft)/(fc+ft)

I have introduced the subdomain expressions fc and ft and defined the function "Disp" (displacement) as follows:

0 0
0.5 0.0005
0 0.001

to define an prescribed displacement of 1 mm.

I'm trying to solve this problem as a time-dependent function but the solution doesn't converge. I have always the Error message:

Error: 8030
Subdomain plot failed.
Failed to compute elasto-plastic variables, try instead the Gauss-point variables.

I'm working in the 35a version of COMSOL and I would like to solve this problem before passing to the most recent version of COMSOL.

If someone has had a similar problem before or know the answer for this kind of problem I would really appreciate her/his help.

Thanks in advance,

Angel


4 Replies Last Post 30.06.2011, 06:49 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10.12.2010, 01:33 GMT-5
Hi

I havent gone far into your model but I'm puzzled about: 1)

sigma_eqv = sqrt(0.5*(sx^2+sy^2+sz^2)+sxy_smsld^2), et

shoulnt it be ?
sigma_eqv = sqrt(0.5*(sx_smsld^2+sy_smsld^2+sz_smsld^2)+sxy_smsld^2), et

then I suspect k is already defined in COMSOl so there might be a little naming mess, try avoid single lettre variables for you own variables, I thendto add a _ or _ik to my variables to clearly identify them


--
Good luck
Ivar
Hi I havent gone far into your model but I'm puzzled about: 1) sigma_eqv = sqrt(0.5*(sx^2+sy^2+sz^2)+sxy_smsld^2), et shoulnt it be ? sigma_eqv = sqrt(0.5*(sx_smsld^2+sy_smsld^2+sz_smsld^2)+sxy_smsld^2), et then I suspect k is already defined in COMSOl so there might be a little naming mess, try avoid single lettre variables for you own variables, I thendto add a _ or _ik to my variables to clearly identify them -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10.12.2010, 05:28 GMT-5
Hi Ivar,

firstly, thank you very much for your answer.

To define sigma_eqv I had previously defined sx, sy and sz as follows:

sx = sx_smsld-sigma_mean,
sy = sy_smsld-sigma_mean,
sz = sz_smsld-sigma_mean, and
sigma_mean = (sx_smsld+sy_smsld+sz_smsld)/3

Apparently the equations are well introduced, anyway I tried to make the change that you suggested me but it doesn't work not either.

The problem is that the model doesn't arrive to converge because of some Error in residual computation.

Besides en the error message they suggest me to try instead the Gauss-point variables, but it doesn't work.

I would really appreciate any kind of help or suggestion that you could give me to solve this problem.

Angel
Hi Ivar, firstly, thank you very much for your answer. To define sigma_eqv I had previously defined sx, sy and sz as follows: sx = sx_smsld-sigma_mean, sy = sy_smsld-sigma_mean, sz = sz_smsld-sigma_mean, and sigma_mean = (sx_smsld+sy_smsld+sz_smsld)/3 Apparently the equations are well introduced, anyway I tried to make the change that you suggested me but it doesn't work not either. The problem is that the model doesn't arrive to converge because of some Error in residual computation. Besides en the error message they suggest me to try instead the Gauss-point variables, but it doesn't work. I would really appreciate any kind of help or suggestion that you could give me to solve this problem. Angel

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10.12.2010, 06:17 GMT-5
Hi

unfotunately I do not have any other ideas to propose like that, the sx looked like strange, but if these are defined variables, now sx is also used internally, perhaps to be on the sure side, try to avoid simple names or personalise your variables fully

--
Good luck
Ivar
Hi unfotunately I do not have any other ideas to propose like that, the sx looked like strange, but if these are defined variables, now sx is also used internally, perhaps to be on the sure side, try to avoid simple names or personalise your variables fully -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30.06.2011, 06:49 GMT-4

Hi Ivar,

firstly, thank you very much for your answer.

To define sigma_eqv I had previously defined sx, sy and sz as follows:

sx = sx_smsld-sigma_mean,
sy = sy_smsld-sigma_mean,
sz = sz_smsld-sigma_mean, and
sigma_mean = (sx_smsld+sy_smsld+sz_smsld)/3

Apparently the equations are well introduced, anyway I tried to make the change that you suggested me but it doesn't work not either.

The problem is that the model doesn't arrive to converge because of some Error in residual computation.

Besides en the error message they suggest me to try instead the Gauss-point variables, but it doesn't work.

I would really appreciate any kind of help or suggestion that you could give me to solve this problem.

Angel


Hi Angel,

I am having a similar problem with the version 4.1. Did you solve the problem? Do you have any idea how to get the variables at Gauss-points? Thanks!


[QUOTE] Hi Ivar, firstly, thank you very much for your answer. To define sigma_eqv I had previously defined sx, sy and sz as follows: sx = sx_smsld-sigma_mean, sy = sy_smsld-sigma_mean, sz = sz_smsld-sigma_mean, and sigma_mean = (sx_smsld+sy_smsld+sz_smsld)/3 Apparently the equations are well introduced, anyway I tried to make the change that you suggested me but it doesn't work not either. The problem is that the model doesn't arrive to converge because of some Error in residual computation. Besides en the error message they suggest me to try instead the Gauss-point variables, but it doesn't work. I would really appreciate any kind of help or suggestion that you could give me to solve this problem. Angel [/QUOTE] Hi Angel, I am having a similar problem with the version 4.1. Did you solve the problem? Do you have any idea how to get the variables at Gauss-points? Thanks!

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.