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.
singularity. error
Posted Aug 4, 2010, 5:04 a.m. EDT Studies & Solvers 10 Replies
Please login with a confirmed email address before reporting spam
Hello,
Does anyboday know why I encounter the below error message in Comsol 3.5a?
Error:
Time 7.86913096623935e-007:
Repeated error test failures. May have reached a
Last time step is not converged.
Thanks.
Does anyboday know why I encounter the below error message in Comsol 3.5a?
Error:
Time 7.86913096623935e-007:
Repeated error test failures. May have reached a
Last time step is not converged.
Thanks.
10 Replies Last Post Jan 2, 2017, 8:19 p.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I have the same error is there any way to over come this error
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Just wondering if anybody was able to resolve this error. I am getting this error in version 4.1 and I am unable to understand from where the singularity is coming from.
Thanks,
Sirisha
Thanks,
Sirisha
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
For me that ressembles typically to a too large step from initial conditions to imposede BC's, check your BC, initial conditions and if you have some large turn on steps, use a smoothing function
--
Good luck
Ivar
For me that ressembles typically to a too large step from initial conditions to imposede BC's, check your BC, initial conditions and if you have some large turn on steps, use a smoothing function
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello,
Does anyboday know why I encounter the below error message in Comsol 3.5a?
Error:
Time 7.86913096623935e-007:
Repeated error test failures. May have reached a
Last time step is not converged.
Thanks.
I am using Comsol 4.2. I got the same error, like this:
"Feature: Time-Dependent Solver 1 (sol1/t1)
Repeated error test failures. May have reached a singularity.
Time : 555.312563858365
Last time step is not converged."
"Feature: Time-Dependent Solver 1 (sol1/t1)
Failed to find consistent initial values.
Out of memory during LU factorization.
Last time step is not converged."
but I found the other way to solve, I reduced the computing time in my model, change the time in STUDY 1 > STEP 1 > TIMES > range , for example, range (0,0.1[h],3[d]) ----> become : range(0,1,3[min])
That is the way I did to solve the problem.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I have the same error but I couldn't reduce the time range.my time range is (0,1[h],60[d]). Is there any other ways that I could solve the problem?
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi guys,
For me this is typically the problem when the initial values are not close enough to the solution, which results in a too large initial step size for the solver and leads to singularity. The problem can thus be resolved by choosing more realistic initial conditions.
Good luck,
Shivam
For me this is typically the problem when the initial values are not close enough to the solution, which results in a too large initial step size for the solver and leads to singularity. The problem can thus be resolved by choosing more realistic initial conditions.
Good luck,
Shivam
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
He,
I get this error:
Repeated error test failures. May have reached a singularity.
Time : 786.5492301366643
Last time step is not converged.
So the problem is not in the beginning???
So I do not think that the problem is that the initial values are not close enough to the solution..
What do you think? And do you have an other solution
Astrid
I get this error:
Repeated error test failures. May have reached a singularity.
Time : 786.5492301366643
Last time step is not converged.
So the problem is not in the beginning???
So I do not think that the problem is that the initial values are not close enough to the solution..
What do you think? And do you have an other solution
Astrid
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Those who are trying to solve singularity by reducing the time:
Like this you don't solve your error. You just did the calculations till u reach singularity.
Rather than this try putting in more boundary conditions.
--
Kalpana Samant
Bioprocess Engineer
Like this you don't solve your error. You just did the calculations till u reach singularity.
Rather than this try putting in more boundary conditions.
--
Kalpana Samant
Bioprocess Engineer
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hello Kalpana,
It may not always be reasonable to add in additional boundary conditions, beyond the default boundary conditions. Each boundary condition that you add should represent the reality that you are trying to model.
The error message being discussed here has no single cause, and can arise from many different situations. Some, but not all, situations that may lead to this type of error include:
- Having a mesh that is too coarse to represent the spatial variations in the field solutions, especially when there are fast transients in the model.
- A boundary condition, or load, which imposes a change in the solution field that is physically incompatible with the governing equations. (For example, think about applying an instantaneous change in a structural displacement boundary condition. This would imply infinite velocity!)
- Having a non-physical nonlinear material definition, or a material property with discontinuities or instantaneous changes in slope. (use a smoothed representation of the material property instead.)
- Check your initial conditions, are they physically possible?
Overall, check the model carefully to make certain that everything makes physical sense. A model which doesn't converge most often is simply set up in some non-physical way. Once you are certain that your model is appropriate, and if you still have problems, you may also find this information helpful:
www.comsol.com/support/knowledgebase/1127/
Of course, there are other reasons beyond just these, but the significant majority of non-convergence issues arise from a some non-physical problem definition.
Best Regards,
Walter
It may not always be reasonable to add in additional boundary conditions, beyond the default boundary conditions. Each boundary condition that you add should represent the reality that you are trying to model.
The error message being discussed here has no single cause, and can arise from many different situations. Some, but not all, situations that may lead to this type of error include:
- Having a mesh that is too coarse to represent the spatial variations in the field solutions, especially when there are fast transients in the model.
- A boundary condition, or load, which imposes a change in the solution field that is physically incompatible with the governing equations. (For example, think about applying an instantaneous change in a structural displacement boundary condition. This would imply infinite velocity!)
- Having a non-physical nonlinear material definition, or a material property with discontinuities or instantaneous changes in slope. (use a smoothed representation of the material property instead.)
- Check your initial conditions, are they physically possible?
Overall, check the model carefully to make certain that everything makes physical sense. A model which doesn't converge most often is simply set up in some non-physical way. Once you are certain that your model is appropriate, and if you still have problems, you may also find this information helpful:
www.comsol.com/support/knowledgebase/1127/
Of course, there are other reasons beyond just these, but the significant majority of non-convergence issues arise from a some non-physical problem definition.
Best Regards,
Walter
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Hi Walter - Just want to let you know that I have been vexed by this error message and your description here got me to finding the problem. I looked harder for something unphysical and found it (a sign error in one of my equations), and now my model is working well. Thanks a lot.
Tom
Tom