I am stuck on this for days. Any help would be appreciated.
The following error has shown:
Repeated error test failures. May have reached a singularity.
Time: 0.00052653794651556622 s.
Merged variables
Attempt to evaluate real square root of negative number.
Function: sqrt
Failed to evaluate variable.
Variable: comp1.mf, Defined as: ((sqrt((0.5*m)/(comp1.T*(pi*Kb)))*comp1.Psat)*(1-Br))*
Failed to evaluate variable.
Hi,
Could you provide more information? I'm afraid without knowing what are your variables it is not possible to answer your question. But anyway it seems at a step during solution one of the variables becomes negative, even though you do not expect it.
Hello Hamidreza,
Thank you for your reply. I am doing laser keyhole welding modelling with the level set method to track the interface of liquid and vapour.Error is in mass flux term as indicated by COMSOL. Recoil pressure is added, vaporization is added too. Please let me know if you want to know something specific.
Thank you.
Still it is too vague to be answered. What is obvious is that one of your variables goes negative somewhere during calculation. Which one, that of course cannot be said by just looking into the equation you mentioned or by knowing what situation you are modelling.
Alright. Thank you Hamidreza. I will look into it.
I got this answer from one of the COMSOL forum. I think I have to make two studies and proceed. and Of course, will check all the parameters one by one.
" Hi, I would get a similar error earlier. As far as I remember, the issue was with the initial conditions that were far from the final solution. To resolve that, solve the heat transfer at first (using u,v,w=0), then use the solution as an initial condition for the combined study. Hope this helps you. Amin, "
Hi,
I read your direct message. Unfortunately I could not open your file because at the time I do not have access to your version of COMSOL, but I opened your file using a text editor and found out something that MAYBE is the source of your problem. Though, not being able to open the file and not being familiar with your discipline I can be wrong. Anyway, here is what I noticed:
It seems in your parameter definitions you have used simple names so that some may interfere with built-in names. In particular "n" which you've used for "absorption of laser" is a built-in name used by COMSOL for surface normal. Check the documentation, and avoid simple names.
(Again) I can be wrong, but it is worth checking it just by simply replacing any simple name (n, P, C, b, etc) with proper ones.
Aside from this, you can define Probes for variables present in your "mf" equation and use Results While Solving feature to see which one and when causes the sqrt's argument become negative. That can give you a clue. For example if it happens right in the beginning then you may need different initial condition.
- Badges
- Science topic