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 The information provided may be out of date.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.
Problem with temperature dependent viscosity
Posted Apr 8, 2011, 4:18 a.m. EDT Heat Transfer & Phase Change, Computational Fluid Dynamics (CFD), Parameters, Variables, & Functions, Results & Visualization, Studies & Solvers Version 3.5a, Version 4.2 7 Replies
Please login with a confirmed email address before reporting spam
As I proceeded further on my model, to make it more physically realistic, I added a temperature dependent viscosity term and enabled convective heat transfer. My viscosity values are based on the VFT relation, and for the type of glass that I am using, I entered a value as follows:
10^(-2.49474+(4099.768/(T[1/K]-544.0237)))[Pa*s]
This causes "negative inside sqrt" error either during calculation of various expressions or the residual term, depending on the type of solver that I use, whenever I enable convective heating.
When I disable convective heat transfer, the solver seems to work fine but gets stuck at certain times and I observe convergence value getting bigger and bigger at those times. When I give a constant value for viscosity (5-10-20-30), the bigger viscosity I give, the less problematic the convergence becomes during solver operation.
Has anyone faced with a similar problem? Any help would be greatly appreciated.
PS: One of my colleagues suggested that I should change my time unit from seconds to minutes or even hours, as glass processes are slow in real life. I do not know how to do that in COMSOL as well.
Please login with a confirmed email address before reporting spam
For the time constant, in 3.5 there is normally a time multiplier, setting it to 60 means default min, 3600 default hours, but I find it easier to use units (at least in V4 you have min, hour, day, year) and just make "t" largeer
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Please login with a confirmed email address before reporting spam
Sorry I cannot find it back, 3.5 is getting old ;) I could not find it either in the doc,. maybe it was/is a application mode specific expression/parameter.
Anyhow what works OK by using t=2[day] or t=36[hour] in both 3.5 and 4.
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Please login with a confirmed email address before reporting spam
Gaelle
Please login with a confirmed email address before reporting spam
Hi
Sorry I cannot find it back, 3.5 is getting old ;) I could not find it either in the doc,. maybe it was/is a application mode specific expression/parameter.
Anyhow what works OK by using t=2[day] or t=36[hour] in both 3.5 and 4.
--
Good luck
Ivar
I am searching this thing and now found it. Thank you so much Ivar.
Best regards.
Please login with a confirmed email address before reporting spam
Please can you explain how do you change the "t" unit ? I have the same error and I don't find how change the time unit. Thank you
Gaelle
I changed time unit by this way, normally (default) time unit is second, now when I want to set up time to solve is minute, hour, day, year... I setup time like this: 3[min] or 5[d] or 8[h]... etc
For example: when the time unit is default, I type time like this (in the quotation mark): Times: "range(0,0.1,120)" second
Now: I can type: Times: "range(0,0.1,2[min])" second.
Hope you can find out what you need.