Unexpected resistance doubling

Clearly defined bug reports and their fixes
Post Reply
jbm
Posts: 2
Joined: 27 Mar 2020, 02:01
Antispam: Yes

Unexpected resistance doubling

Post by jbm »

I am trying to compute electrical resistance with ElmerSolver 8.4. I always get twice the resistance I would expect, e.g. for a simple block of 1 x 1 x 1 material of conductivity 1 with a potential difference applied between the left and right faces I would expect a resistance of 1 according to R=length/(area*conductivity). StatCurrentSolver reports a resistance of 2.

In elmerfem/fem/tests/ComponentResistance/ a component of length 1, width 0.2 and conductivity 1.0e3 is reference normed to have a resistance of 0.01, twice what I would expect from 1/(0.2*1.0e3) = 0.005. So apparently this is the desired behaviour but I do not understand why.
Attachments
case2.sif
(1.68 KiB) Downloaded 140 times
blocks2.grd
(606 Bytes) Downloaded 132 times
case2.log
(5.58 KiB) Downloaded 133 times
kevinarden
Posts: 1210
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Unexpected resistance doubling

Post by kevinarden »

If I turn on
Calculate Volume Current = Logical True
in the solver section I get
Variables in columns of matrix: ./f.dat
1: res: total joule heating
2: res: effective resistance
3: res: comp 1: electric resistance potential
4: component 1: res: electric resistance

and 2, the effective resistance is 1.0
but 4 electrical resistance is 2.0
kevinarden
Posts: 1210
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Unexpected resistance doubling

Post by kevinarden »

I = V/R holds when calculating the current so internally the resistance must be 1, or it is using effective resistance to calculate current.
jbm
Posts: 2
Joined: 27 Mar 2020, 02:01
Antispam: Yes

Re: Unexpected resistance doubling

Post by jbm »

Thanks for looking at this Kevin. I also get that when switching calculate volume current on, unfortunately this only further confuses me.
Post Reply