Discussion Forum

Posted: 1 decade ago 24 giu 2012, 08:32 GMT-4

Hi my first suggestion (without accessing COMSOL, I'm not at work) would be to say: try frequency domainsolver and set the frequency in the solver entry, then replace your "electric potential" by a "terminal" node, you should then get lumped parameters out (I always think of Terminal = average value identical over the full boundary = connection to external worls (cable etc) and Electric potential = distributed V(x,y,z,t...) potential that might well not be an average scalar value, hence no lumped impedance values pre-cooked by COMSOL. -- Good luck Ivar

Posted: 1 decade ago 24 giu 2012, 13:35 GMT-4

Ivar, thank you very much! Okay, i replace "electric potential" by a "Port" node, chose the "Forced voltage" as input property, set point "Use port as input" and type Vin = "1" V Port voltage. My task is plot frequency dependence of the phase and absolute value of the impedance. In "Solver Parameters" window I select "Time-harmonic, electric currents" as analysist type, Solver - "Parametric", Parameter name - nu_emqvw, value - range(100,900/2,1000) (I find nu_emqvw in window "Application Scalar Variables", Description - "Frequency"). Meshing, Solving. I can plot addmitance matrix versus frequency, but how i can plot phase of the admittance (or impedance)?

Posted: 1 decade ago 24 giu 2012, 16:25 GMT-4

Hi inded I believe your are right there, I understand you are in 3.5 that is getting a bit old for me and the names have changed with v4 too. Normally the amplitude and phase of the phasor is obtained by "real(value)" and "atan2(imag(value),real(value))" -- Good luck Ivar

Posted: 1 decade ago 25 giu 2012, 01:20 GMT-4

Appeared the following question: How I can plot dependence of current thru a plane boundary versus variable parametr (frequency). I think i must integrate the "Total current density" over the poundary, and plot obtained value versus frequency. How i can do it? P.S. Sorry for my bad English...

Posted: 1 decade ago 25 giu 2012, 02:38 GMT-4

Hi (In V4 because I have forgotten how it was in v3.5 there was a derived or global GUI window somewhere ;( use the Results Derived Variables - Integration to integrate your current density over your boundary. If you have solved your problem with a frequency domain sweep then you will get automatically a table out per frequency step, and then you can plot this -- Good luck Ivar

Posted: 1 decade ago 25 giu 2012, 16:46 GMT-4

Dear Ivar! How i can plot expression, which consist of voltage on some boundary and integral of current density over this boundary? E.g. V/[integral of (ec.Jz)]?

Posted: 1 decade ago 26 giu 2012, 01:45 GMT-4

Hi in V4 you have the Results Derived Variables to plot values into tables then you can reformat them at will. Other way you Definition - Variables (set up variables) then solve and then refer to the variables. You construct your variables with the coupling operators you define in the Definition - Couplings node -- Good luck Ivar

Posted: 1 decade ago 26 giu 2012, 18:00 GMT-4

Thanks, everything worked out just great!

Attachments:

• simple model AC impedance.mph

Posted: 1 decade ago 27 giu 2012, 07:51 GMT-4

Hi but first of all is epsilon_r = 0 really an acceptable material value, I'm not sure Second you have also the (pre-cooked for you by COMSOl with the Terminal BC node) the admittance (Derived Values - Global Evaluation - terminal admittance ec.Y11 which is a complex value (a phasor, if you take the absolute value you find back your Z impedance, but if you take the phase value atan2(imag(ec.Y11),real(ec.Y11)) you will also see the phase lag -- Good luck Ivar

Posted: 1 decade ago 28 giu 2012, 05:03 GMT-4

Hellow! I wanted to learn how modeling some structures in AC/DC Comsol, therefore i define epsilon_r = 0 for simplicity the model. And I know, that Comsol hase many friendly built-in functions (such as admittance ), but I wanted to study Comsol. Now I want define some property of the material as a function of the spatial coordinate. Is it possible? I found that if i define some property as "Anisotropic" (for example Electrical conductivity), then I can define properties in 9 (for 2D model) areas. But it is very restricted method. For example, how define Electrical conductivity as a sin(spatial coordinate*koefficient)?

Posted: 1 decade ago 28 giu 2012, 05:22 GMT-4

Hi at least then use epsilon_r = 1 and not "0" I do not believe that is physical ;) When you add a value into a place in the BC nodes, COMSOL accepts a scalar, or a function and this can well be a function of the spatial or time coordinates. So if you enter a temperature T(x,y,z,t) as a complete field and you define the variations of T in an interpolation function, then COMSOL will apply that to your BC The for the materials, you can indeed use the full tensor representation with full or partial anisotropy, just as well as you can propose interpolation functions to have physical parameters depend on T, p, B, E ... but you make your models more or less non-linear hence it takes time to solve -- Good luck Ivar

Posted: 1 decade ago 16 lug 2012, 15:16 GMT-4

[QUOTE] Ivar, thank you very much! Okay, i replace "electric potential" by a "Port" node, chose the "Forced voltage" as input property, set point "Use port as input" and type Vin = "1" V Port voltage. My task is plot frequency dependence of the phase and absolute value of the impedance. In "Solver Parameters" window I select "Time-harmonic, electric currents" as analysist type, Solver - "Parametric", Parameter name - nu_emqvw, value - range(100,900/2,1000) (I find nu_emqvw in window "Application Scalar Variables", Description - "Frequency"). Meshing, Solving. I can plot addmitance matrix versus frequency, but how i can plot phase of the admittance (or impedance)? [/QUOTE] Hello Could u please help me with finding the POrt node under electric! I am trying to input an ac voltage in piezo physics and preety much doing the same thing as you...but just cannot find...the port node and lumped parameters everyone is talking about. also i am using 4.2a ...cant open your model for reference too! thanks.megha

Posted: 1 decade ago 16 lug 2012, 15:22 GMT-4

Hi it's Terminal in v4 (ES or EC) in a harmonic solver case (frequency domain) you do only define the amplitude, and the phaseif applicable in th BC node, then the frequency or frequnecy range in the solver node -- Good luck Ivar

Posted: 1 decade ago 11 mar 2014, 12:22 GMT-4

Hi all, i want calculate the capacitance between the two faces (up and down), how can I do??? please help me!

Attachments:

• prova semplice.mph