Jump to content


Photo
- - - - -

Electrical resistivity in finite elements


  • Please log in to reply
11 replies to this topic

#1 baltoche

baltoche

    Quark

  • Members
  • 24 posts

Posted 15 February 2017 - 08:43 PM

Dear All,

 

Suppose I've got a medium (like a soil) of a homogeneous and constant resistivity Rho (in Ohm.km). This medium is put into a big cube of 1km side, then giving a km3 of my medium.

 

Now suppose I want to model my cube by an electrical circuit with resistors at each side of the cube. May I state that each resistor will be Rhox4 then ?

 

No suppose I divide my km3 into equal 8 cubes of 500m side each and suppose I still want to model this arrangement by electrical resistors. What should be the value of each resistor/side then if I need that the whole set still corresponds to a resistivity of Rho ? I propose Rhox9/2. Am I right ?

 

For these who know Matlab, can Matlab simulate this kind of 3-D electric circuit with cubes whose sides are resistors ? Is there any special module to perform this function ?

 

Thanks to All.

 

 


  • 0

#2 Bender

Bender

    Molecule

  • Senior Members
  • 627 posts

Posted 15 February 2017 - 09:00 PM

Of course Matlab can do that. There is even a Finite Element toolbox, or you can compose the matrices yourself.


  • 0

#3 baltoche

baltoche

    Quark

  • Members
  • 24 posts

Posted 16 February 2017 - 01:32 PM

Thanks for your reply. Is there any special module to access to this functionality ? I have the basic licence only.

 

Regards.


  • 0

#4 Bender

Bender

    Molecule

  • Senior Members
  • 627 posts

Posted 16 February 2017 - 02:21 PM

You can switch to Octave-Forge, which is free and has these toolboxes available, but I have no experience with those. Octave uses mostly the same code as Matlab.


  • 0

#5 baltoche

baltoche

    Quark

  • Members
  • 24 posts

Posted 16 February 2017 - 08:07 PM

I'll try this, thanks. But is Octave-Forge somehow related to Matlab or not ?


  • 0

#6 Bender

Bender

    Molecule

  • Senior Members
  • 627 posts

Posted 16 February 2017 - 08:11 PM

Only that it uses the largely same code.


  • 0

#7 baltoche

baltoche

    Quark

  • Members
  • 24 posts

Posted 16 February 2017 - 10:59 PM

Oh I see, thanks !

Do you personally use these tools, and especially the circuit simulation one ?


  • 0

#8 Bender

Bender

    Molecule

  • Senior Members
  • 627 posts

Posted 16 February 2017 - 11:56 PM

No. I have written an m-file for structural mechanics problems, but that was mainly for didactic purposes.

 

Of course, you could also use an actual finite element program, such as Comsol or NX, if you have access to such. Probably a lot easier to get into.


  • 0

#9 baltoche

baltoche

    Quark

  • Members
  • 24 posts

Posted 17 February 2017 - 09:05 AM

Do you know if Comsol also deals with electrical fields ? Is it a freeware software ?


  • 0

#10 Bender

Bender

    Molecule

  • Senior Members
  • 627 posts

Posted 17 February 2017 - 11:40 AM

It deals with just about anything you want. It is also very expensive (like Matlab). I only have experience with modelling magnetic fields with it.

 

If you want open source: I have heard very good things about Elmer, but I have never used it. From a brief look years ago, I had the impression that it is slightly harder to get started with it.

 

Do you have any experience with finite element modelling? A warning: do not take lightly to getting into it, as getting decent results is not straightforward. Even results that look good can be rubbish.


Edited by Bender, 17 February 2017 - 11:41 AM.

  • 0

#11 baltoche

baltoche

    Quark

  • Members
  • 24 posts

Posted 20 February 2017 - 04:06 PM

Well caught, thank you very much for your support from the beginning. I'll not be involved myself into finite elements, but an engineering student will dot it for me. I've got a licence of CST, do you know this tool ?

One last question please : have you got any opinion on the resistor value of my cube edges ?

 

Kind Regards.


  • 0

#12 Bender

Bender

    Molecule

  • Senior Members
  • 627 posts

Posted 21 February 2017 - 09:16 AM

I don't know CST.

 

Just one remark for the resistor values: it is not the resistivity that changes, but the surface area. At first glance replacing a single resistor with four resistors of one fourth the resistance should work, at least in the principal directions. I don't know how accurate the results will be in a random direction.


  • 0




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users