Menu

Discussion Forum

New Discussion
Filter

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 The information provided may be out of date.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.

EIS - impedance measurements with very low conductivity

Posted Jan 27, 2012, 6:34 a.m. EST Low-Frequency Electromagnetics, Results & Visualization, Studies & Solvers Version 4.0a, Version 4.2a 5 Replies

Ebia
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

I've bee working for a while on a EIS model, so models of electrical impedance spectroscopy on various shape of electrodes (2D model properly setting the d [m] )

(Terminal 1 and 2 in voltage configuration, frequency analysis)

It happens that at a very low frequency the real part of the impedance change into negative....

I can't find a physical reason to obtain this so I suppose it is something related to the precision of the calculation

I did a lot of tests and I found out that randomly, a very low conductivity setting to an interface or to a materials results in a negative real term of impedance (1/ec.Y11).

...? a bug? a not proper numerical precision?

help please
best

5 Replies Last Post Jun 22, 2012, 9:53 a.m. EDT
Edgar Kaiser
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 27, 2012, 6:55 a.m. EST
How small are the numbers? Is it approaching machine precision?
How small are the numbers? Is it approaching machine precision?
Ebia
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 27, 2012, 8:06 a.m. EST

We expect to obtain Real Z value under 10e+8 Ohm
we deal with conductivity around 10e-10 S/m
( this values are very common in the EIS field )

we check the machine precision on Matlab and we found a eps = 2.2 e-16


but actually, depending on the complexity of the model, we obtained very consistent values also as 1e+30 ( real Z).

Is it possible that we go under the machine precision during the calculation from conductivity -> admittance?

is there any section in the documentation , helping us to understand the mathematical model used by Comsol to reach the result?

thanks
We expect to obtain Real Z value under 10e+8 Ohm we deal with conductivity around 10e-10 S/m ( this values are very common in the EIS field ) we check the machine precision on Matlab and we found a eps = 2.2 e-16 but actually, depending on the complexity of the model, we obtained very consistent values also as 1e+30 ( real Z). Is it possible that we go under the machine precision during the calculation from conductivity -> admittance? is there any section in the documentation , helping us to understand the mathematical model used by Comsol to reach the result? thanks
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 27, 2012, 8:57 a.m. EST
Hi

If you type in Parameter epss = "eps" (without the "", you get the COMSOl precision, for me 2.220E-16. but you should not forget that most physics imply second order formulas and differences and sqrt(eps) = 1.49E-8

And eps is valid for the difference from "1" if you are scaled far from "1" the smallest difference is much much greater

Check any book on binary representation and numerical precisions

--
Good luck
Ivar
Hi If you type in Parameter epss = "eps" (without the "", you get the COMSOl precision, for me 2.220E-16. but you should not forget that most physics imply second order formulas and differences and sqrt(eps) = 1.49E-8 And eps is valid for the difference from "1" if you are scaled far from "1" the smallest difference is much much greater Check any book on binary representation and numerical precisions -- Good luck Ivar
Parmender Singh
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 28, 2012, 4:31 a.m. EDT
Hello

Can you send me your model . i am also working on Battery analysis though magnetic filed. so i have no idea how to couple AC/Dc module with battery model. Please help me
Hello Can you send me your model . i am also working on Battery analysis though magnetic filed. so i have no idea how to couple AC/Dc module with battery model. Please help me
Richie Guo
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 22, 2012, 9:53 a.m. EDT
Did you recieve the Model？
Did you recieve the Model？

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.

Suggested Content