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 The information provided may be out of date.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.

Non-Isothermal flow module issues

Please login with a confirmed email address before reporting spam

Hi, I am trying to model a single channel of a typical pressurized water reactor, considering three physics, the point kinetics model for the calculation of energy generated inside the fuel pellet, the heat transfer, and the turbulent flow of the coolant water.
Please take a look at the attached two files. In the first file, I used the Non-Isothermal flow (nitf) k-epsilon model, and implemented the point kinetics model by the Global node, and the it works! However, if I change the heat transport turbulence model from "Kays-Crawford" to "Extended Kays-Crawford" option, it does not work!
The only difference between the first and second file, is that, I put the ODE equation outside of the Non-Isothermal flow (nitf) module, and it could not converge!

Could anyone explain to me why the above things happen and what mistakes I have made there?
Thanks.

Lei


6 Replies Last Post 29 janv. 2014, 17:20 UTC−5
Luke Gritter Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 janv. 2014, 10:33 UTC−5
Lei,

When you use the Global ODEs and DAEs interface, the software generates slightly different default solver settings compared to when you add the global equations in the Non-Isothermal Flow interface. For the case with the Global ODEs and DAEs interface, the error estimation (under Advanced on the Time-Dependent Solver node in the solver sequence) is set to "Include algebraic", while in the other case it is set to "Exclude algebraic". This solver setting is the cause of the different behavior of the time-dependent solver in the two models.

--
Luke Gritter
AltaSim Technologies
Lei, When you use the Global ODEs and DAEs interface, the software generates slightly different default solver settings compared to when you add the global equations in the Non-Isothermal Flow interface. For the case with the Global ODEs and DAEs interface, the error estimation (under Advanced on the Time-Dependent Solver node in the solver sequence) is set to "Include algebraic", while in the other case it is set to "Exclude algebraic". This solver setting is the cause of the different behavior of the time-dependent solver in the two models. -- Luke Gritter AltaSim Technologies

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 janv. 2014, 11:11 UTC−5
Hi, Luke:
Thank you for your reply. I tried changing the Advanced option from "Include algebraic" to "Exclude algebraic", and it is working the same way as using the Global node.
Next, I want to add a temperature feedback relation into the system, in the sense that if the temperature rises due to the power (heat source) increase, it will give certain feedback reactivity. In my model, I take average temperature over the entire domain to form a feedback reactivity (defined as a variable), then use the variable to get the new reactivity "rho" as input condition for calculation of the power, generating new heat source, thus influencing the heat transfer solutions. However, the coupled system does not seem to compute.
The error message is:
Repeated error test failures. May have reached a singularity.
Time : 0.06456702759609664
Last time step is not converged.
See the attached file.
Is there anything that I messed up here? Could you please help me with that?
Thanks.

Lei
Hi, Luke: Thank you for your reply. I tried changing the Advanced option from "Include algebraic" to "Exclude algebraic", and it is working the same way as using the Global node. Next, I want to add a temperature feedback relation into the system, in the sense that if the temperature rises due to the power (heat source) increase, it will give certain feedback reactivity. In my model, I take average temperature over the entire domain to form a feedback reactivity (defined as a variable), then use the variable to get the new reactivity "rho" as input condition for calculation of the power, generating new heat source, thus influencing the heat transfer solutions. However, the coupled system does not seem to compute. The error message is: Repeated error test failures. May have reached a singularity. Time : 0.06456702759609664 Last time step is not converged. See the attached file. Is there anything that I messed up here? Could you please help me with that? Thanks. Lei


Luke Gritter Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 janv. 2014, 15:44 UTC−5
Lei,

Your feedback term is driving your heat source to infinity, so there is clearly something non-physical in your implementation of the physics. This is the cause of the convergence failure.

--
Luke Gritter
AltaSim Technologies
Lei, Your feedback term is driving your heat source to infinity, so there is clearly something non-physical in your implementation of the physics. This is the cause of the convergence failure. -- Luke Gritter AltaSim Technologies

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 janv. 2014, 15:48 UTC−5
Luke,
I see that problem now. It should be the opposite sign. OK, it is working now. Thanks for pointing it out.
By the way, do you know the reason for the error from changing the heat transport turbulence model from "Kays-Crawford" to "Extended Kays-Crawford" option?
Lei
Luke, I see that problem now. It should be the opposite sign. OK, it is working now. Thanks for pointing it out. By the way, do you know the reason for the error from changing the heat transport turbulence model from "Kays-Crawford" to "Extended Kays-Crawford" option? Lei

Luke Gritter Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 janv. 2014, 16:51 UTC−5
Lei,

When you use the Extended Kays-Crawford model for the turbulent Prandtl number, you must manually define a "Reynolds number at infinity" on the Fluid node. Make sure that you have entered an appropriate value for this Re - if this is left at the default value of zero, it will cause an error.

--
Luke Gritter
AltaSim Technologies
Lei, When you use the Extended Kays-Crawford model for the turbulent Prandtl number, you must manually define a "Reynolds number at infinity" on the Fluid node. Make sure that you have entered an appropriate value for this Re - if this is left at the default value of zero, it will cause an error. -- Luke Gritter AltaSim Technologies

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 janv. 2014, 17:20 UTC−5
Luke,
You are right! It definitely needs to be set to some reasonable value.
Thank you for your help!
Lei
Luke, You are right! It definitely needs to be set to some reasonable value. Thank you for your help! Lei

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.