[Home][2014 Index] Re: [TCML] Yet another TC modeling program [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] Yet another TC modeling program




After some improvements,

 http://abelian.org/lcnetgen

I implemented the fast multipole method for the capacitance
calculation.   This allows large numbers of tiles to be used.

Isolated unit sphere: 111.32 pF using 30,000 tiles
                      111.272 pF using 1,000,000 tiles
                      111.269 pF using 4,000,000 tiles
(correct value is close to 111.26501 pF)

Isolated unit cube: 73.54 pF using 30,000 tiles
                    73.508 pF using 720,000 tiles
(correct value is close to 73.510898 pF)

The 4e6 tile sphere takes about an hour using 32 cores
of Xeon E5-2680.  Fortunately for practical use 10k to 50k
tiles is plenty for around 1% precision and runs in a few
minutes on a domestic PC.  Quad core and 8Gbyte RAM is fine.

Infinite ground plane is partially supported (E-field only).

Calculates response to incident E and H fields.

Example 5 shows the complicated spectrum of a LF loading coil
made from seven nested bucket coils.  This is a difficult one
because the coils are so close together.

I'm looking out for interesting systems to model, especially
where accurate measurements are available for comparison.
--
Paul Nicholson
--
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla