Discussion Forum

Posted: 1 decade ago 4 mars 2013, 05:43 UTC−5

Hi you should rather use "t" as argument no ? "t" is the COMSOL internal name for the time in seconds The segregated solver error means that COMSOl has split the dependent variables but your links makes it incoherent, it proposes that you group them manually: select the first group, open the list of variables "+" and add those missing, then delete the second, or try adding a subnode: Fully Coupled. Note that a transient solver is for "transients, and not stationary cases. If you have a steady state AC signal the Frequency domain solvers are better, but as you have DC + AC you need a "prestressed" analysis, with first a stationary solver for the bias of 50VDC thena harmonic node for the Vac value. Check the doc and the solver cases for harmonic stationary ... (I', not by ma WS so I cannot check just now ;) -- Good luck Ivar

Posted: 1 decade ago 4 mars 2013, 08:59 UTC−5

[QUOTE] Hi you should rather use "t" as argument no ? "t" is the COMSOL internal name for the time in seconds The segregated solver error means that COMSOl has split the dependent variables but your links makes it incoherent, it proposes that you group them manually: select the first group, open the list of variables "+" and add those missing, then delete the second, or try adding a subnode: Fully Coupled. Note that a transient solver is for "transients, and not stationary cases. If you have a steady state AC signal the Frequency domain solvers are better, but as you have DC + AC you need a "prestressed" analysis, with first a stationary solver for the bias of 50VDC thena harmonic node for the Vac value. Check the doc and the solver cases for harmonic stationary ... (I', not by ma WS so I cannot check just now ;) -- Good luck Ivar [/QUOTE] Thank you for your kind replay! I think I have made the Vac, but I got into trouble in the simulation just as the figure error shows.It tells me that: "Failed to evaluate variable Jacobian. —Variable:t —Geometry:1; —Boundary:10 —Feature:Stationary Solver 1(sol1/s1); Error:Failed to evaluate variable Jacobian " Could you give me some directions about this please? And actually I just choose the stationary solver and I think that you're right. But I'm sorry that I don't understand your suggestion about the solver, could you describe it a little more detailedly please?Thank you!

Attachments:

• error.jpg

Posted: 1 decade ago 4 mars 2013, 14:27 UTC−5

Hi first, in a stationary case, use rather something else than "t" as it's a reserved variable for COMSOL, except if you are preparing a t=0 or t=cte value for a later transient solver for t>0 or t>constant The Jacobian error means that the definition you have given to "t" is not continuous, and the solver needs continuos fields to be able to derive them for the solving process -- Good luck Ivar

Posted: 1 decade ago 5 mars 2013, 01:41 UTC−5

Hi, Ivar Thank you for your reply! You said that the Jacobian error means that the definition you have given to "t" is not continuous, I didn't find the place where I make "t" continuous.In the Global definition,I define the Waveform and Analytic function,and I made the Arguments be "t". I use this way to add the harmonics, and also get the error:failed to evaluate the Jacobian .So could you tell me how to make it continuous please? Thank you! Best regards

Posted: 1 decade ago 5 mars 2013, 02:37 UTC−5

Hi the variables in the function definitions are dummy names, not linked to anything from the model. Then it must be that your functions, as defined, are not continuous -- Good luck Ivar

Posted: 1 decade ago 5 mars 2013, 08:16 UTC−5

Thank you very much for your reply!