XANSYS Message: 92710 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Rate item:

Subject: Re: [EMAG] Convergence problem in
Author: Krunal D Kansara
Date: 2008-04-07 03:13:05

Hi mike

Thanks for the reply.

Krunal Kansara
SDDC, EBG,
Larsen and Toubro Limited
Mumbai

>>> "Yaksh, Mike" 4/6/2008 6:32 PM >>>
Hi Krunal

If all materials are free space, I would use the 97. It has better
options for using voltage fed system, including the circ124. This can
be used for static, harmonic or transient.

On item 3. below it is correct to have the flux parallel condition at
the surface where the conductor starts and stop. But in most bus
problems, or problems without iron, the other surfaces must be remote in
order to use the flux parallel. Another advantage of the 97 is that the
111 has MVP dofs and it can serve as an infinite boundary element.

On this current issue, turn that convergence off and look at the
solution. If curr is the problem, then the following will turn it off.
Cnvt,curr,1,.01
Cnvt,curr,-1

Mike Yaksh
NAC International
Norcross GA

-----Original Message-----
Krunal D Kansara [krunaldk_at_lntebg.com]


>>With the 117 keyopt(1)=1, you get the time integrated volt, which
means
>>that when you want to apply a real volt, you need a factor of omega in
>>the boundary condition.
I am sorry but i haven't understood the omega factor which you have
mentioned. Do i really need to multiply something to the boundary
condition in keyopt(1)=1.

Solid117 keyopt(1)=1 is transient or harmonic case, so i can understand
that it requires the iterations. But i have also observed that with
Keyopt(1)=5, which is not time integrated voltage, It requires the
iterations and in this case also it never converges with program
calculated convergence criterion.

In my problem, i have an air Keyopt(1)=0 and a conductor which ends at
boundary and voltage fed option with keyopt(1)=5 static analysis.
1. i don't have any permeable material so no BH data.
2. Type of analysis is static only with voltage fed option
3. I am using flux parellel on all six outer area of a cubical air box
with conductor ends are at boundary of air box.
4.I am not using any other options.

I don't understand why it requires iterations and why with program
calculated convergence criterion it never converges.
Posts possibly associated with message #92710AuthorDateScore
92654[EMAG] Convergence problem in magnetostatic analysisKrunal D Kansara 2008/04/03 
92657Re: [EMAG] Convergence problem in magnetostatic analysiswithYaksh, Michael C 2008/04/03 
92677Re: [EMAG] Convergence problem in magnetostaticKrunal D Kansara 2008/04/04 
92705Re: [EMAG] Convergence problem in magnetostaticanalysiswithMike Yaksh2008/04/06 
92710Re: [EMAG] Convergence problem inKrunal D Kansara 2008/04/07