search

ElmerCSC / elmerfem

Official git repository of Elmer FEM software
http://www.elmerfem.org
Other
1.18k stars 317 forks source link

r_component is by localC larger than it should be #392

Closed jvencels closed 1 year ago

jvencels commented 1 year ago

Hi @juharu, There were changes a month ago that affected circuit outputs.

The source is fem/src/modules/CircuitsAndDynamics.F90. Coil conductance is not being multiplied by local conductivity: https://github.com/ElmerCSC/elmerfem/commit/2373af0e6bf051ab49f97be759391a64a9d4374b#r108302956

juharu commented 1 year ago

Hi Juris

Thanks for looking into this!

Could you be a bit more specfic, where do you see the problem exactly (code and / or a test case ? ), harmonic vs transient ? For example, the code shows changes like

cmplx_val = IP % s(t) detJ localC SUM (gradv Wbasis(j,:)) Comp % VoltageFactor cmplx_val = IP % s(t) detJ SUM (gradv Wbasis(j,:)) * Comp % VoltageFactor END IF END IF

IF ( .NOT. SkinBc) cmplx_val = cmplx_val localC + im Omega cmplx_val localP

E.g. the last line multiplies by "localC" removed from the line above ? Can't directly see the trouble, there are a lot of changes tough, so probably I'm missing something...

Thanks!

br, Juha

From: "Juris Vencels" @.> To: "ElmerCSC" @.> Cc: "juharu" @.>, "Mention" @.> Sent: Monday, 10 April, 2023 17:26:36 Subject: [ElmerCSC/elmerfem] r_component is by localC larger than it should be (Issue #392)

Hi [ https://github.com/juharu | @juharu ] , There were changes a month ago that affected circuit outputs.

The source is fem/src/modules/CircuitsAndDynamics.F90. Coil conductance is not being multiplied by local conductivity: [ https://github.com/ElmerCSC/elmerfem/commit/2373af0e6bf051ab49f97be759391a64a9d4374b#r108302956 | 2373af0#r108302956 ]

— Reply to this email directly, [ https://github.com/ElmerCSC/elmerfem/issues/392 | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ACTOMSSQZT4I7MGXWN2IGVTXAQKBZANCNFSM6AAAAAAWZBI5JQ | unsubscribe ] . You are receiving this because you were mentioned. Message ID: @.***>

jvencels commented 1 year ago

Hi Juha,

The test case with massive coils attached. The issue seems to be present for massive coils only.

See /1962/dat/1962.dat.csv, r_component(1). The value we obtain with the older Elmer version is 4.86E-006. With the current version, we get 149.

circuits_harmonic_massive.zip

juharu commented 1 year ago

Thanks,

fix in "devel" ?

From: "Juris Vencels" @.> To: "ElmerCSC" @.> Cc: "juharu" @.>, "Mention" @.> Sent: Tuesday, 11 April, 2023 10:47:20 Subject: Re: [ElmerCSC/elmerfem] r_component is by localC larger than it should be (Issue #392)

Hi Juha,

The test case with massive coils attached. The issue seems to be present for massive coils only.

See /1962/dat/1962.dat.csv, r_component(1). The value we obtain with the older Elmer version is 4.86E-006. With the current version, we get 149.

[ https://github.com/ElmerCSC/elmerfem/files/11197926/circuits_harmonic_massive.zip | circuits_harmonic_massive.zip ]

— Reply to this email directly, [ https://github.com/ElmerCSC/elmerfem/issues/392#issuecomment-1502850536 | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ACTOMSSHCXCB5BRH4DP2SKLXAUEARANCNFSM6AAAAAAWZBI5JQ | unsubscribe ] . You are receiving this because you were mentioned. Message ID: @.***>