Discussion Forum

Hello Harkirat Bansal

Your Discussion has gone 30 days without a reply. If you still need help with COMSOL and have an on-subscription license, please visit our Support Center for help.

If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.

I get that too, has anyone met it/found out how to get around it?

I am also finding this error popping up.

I am consistently having this problem now and it is becoming an unsolvable issue with my model.

I am running Non-Isothermal Flow physics with 2 point-ODE's and a domain ODE. The ODE's and NITF physics are coupled through Temperature, T. The ODE's are fed T and spit out a value used in the NITF physics for a heat source term. The problem is modelling simple neutron kinetics in a liquid reactor.

NITF will run correctly when the ODE's are off, but consistently runs into the error noted above when the ODE's are on. The error most often occurs in a domain in which the ODE's are not solving in and is preceded by a small area suddenly having a large Temp increase (>600K) or similar temperature decrease.

If there is anybody out there that knows why this error is occurring, that would be great. I have tried re-meshing to finer and finer mesh, increasing my discretization, and reducing the time steps to extremely small size (0.001s). Nothing has worked.

Thank you for any help!

Edit: This is for a time-dependent problem.

Hi

are you sure your ODE enters gently the game, and do not pop in with a step function, giving issues for the solver to get to converge. And are all your BC "smooth"

--
Good luck
Ivar

Hi Ivar,

I have checked and double checked my parameters and everything looks alright.

I think I have solved the problem, however. To fix the issue I was having, I went into the Fully Coupled node within the Time Dependent Solver 1 node, within the Solver 1 node in Solver Configurations.

Within the Method and Termination Tab in the Fully Coupled node there is the option to change the nonlinear method. I changed this from Constant (Newton) to Automatic (Newton). This has solved the solver crashing issue. This allows for COMSOL to decide on a varying damping factor, if I understand the documentation correctly.

I tried Automatic Highly Nonlinear but the solver was not happy with that in Time Dependent mode so I have stuck with Automatic and the model is running smoothly now.

I hope this helps anybody else having this error.

Hi all
well, i have actually got this very same error when building the model from the model library (see attached - the PDF from comsol site),

and tried all the above suggestions... nothing so far worked.
Anyone?

I am getting this error as well through simulating a boundary condition with a periodic piecewise function shaped as a step function. The step function is smoothed a little bit but I'm trying to keep the wave as square as possible to emulate real life conditions.

Ivar, you specifically asked about this in the comment above. How badly with prevent a solution for my model. Do I need to make the function more smooth? Could modify the solver settings to accomodate for large and sudden changes in the boundary conditions?

Thanks,
Trevor

Hi

are you sure your ODE enters gently the game, and do not pop in with a step function, giving issues for the solver to get to converge. And are all your BC "smooth"

--
Good luck
Ivar

Hi Ivar,

could you please explain a bit more on how an ODE should enter gently without a step function? My model is a similar case but using with PDE on comsol 4.4. Thank you.

Regards,
Edmund

Dear sir..

Thanks for your help .. and attention

i have got a problem with my model while running as the solve doesn't recover
(Recoverable error in residual, but solver could not recover.
Time : 0.03802165133617842
Last time step is not converged.)

this is the error i found .. please help me to recover this error..

Also i need to add a road profile to my model ( road vehicle simulation) , but i couldn't find the way to add it by the way i tried to add it as a waveform but i think its wrong..

please help me to add this function to the vehicle model as a road excitation

best regards