| T O P I C R E V I E W |
| tomed |
Posted - Mar 29 2015 : 16:17:44
Hello Enrico,
Please is there any rules to correctly define the reduced order model (-r) before running the simulation?
In fact, this number gives an impact to the results regarding R,L values.
Thank you for your help and continuous support
Regards,
Tomed |
| 3 L A T E S T R E P L I E S (Newest First) |
| Enrico |
Posted - Apr 01 2015 : 09:22:46
quote:
May be should make sense to launch a parametric study making the reduced order model changing to find the more accurate order???
Yes but this depends on your case. If R and L are slowly changing, a low order model may be accurate enough, viceversa if R and L have sharp changes you need a higher order ROM. So since this eventually depends on your model, it goes case by case.
On the ROM, yes it is Arnoldi. For a concise digression on the method, you may refer to L. M. Silveira, M. Kamon, J. White, "Efficient Reduced-Order Modeling of Frequency-Dependent Coupling Inductances associated with 3-D Interconnect Structures", IEEE Transactions on Components, Packaging and Manufacturing Technology, Vol. 19, Issue 2, May 1996
Best Regards,
Enrico
|
| tomed |
Posted - Apr 01 2015 : 08:40:30
Hello Enrico,
One more time thank you for your support.
May be should make sense to launch a parametric study making the reduced order model changing to find the more accurate order???
Anyway, still have some questions:
is there any link between this the order model and Pade approximation order or Arnoldi's one?
Thank you for your help and continuous support
Regards,
Tomed |
| Enrico |
Posted - Mar 31 2015 : 20:07:40
Of course this is impacting the result. However defining the ROM order a-priori is not so easy. I would suggest the approach at page 21 of the user's guide to determine the accuracy of the reduced-order model over a reference structure.
Then for similar structures you can apply the same order, as a rule of thumb, for having equivalent accuracy.
Best Regards,
Enrico
|
|
|
