This is a look at your earlier message at 9:23:43 AM.
f=: 13 : '+/-: %>: +:i.y'
f 100000
3.36911
5!:4 <'f'
-- [:
+- / --- +
│
--+ -- [:
│ +- -:
L-----+ -- [:
│ +- %
L----+ -- [:
│ +- >:
L----+ -- [:
L----+- +:
L- i.
A pretty tree. However you have left me with your 11:45 PM post unless you
provide some small numerical examples of the matrices and the inverses you
are looking for.
Linda
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Donald Kelly
Sent: Wednesday, January 09, 2013 11:45 PM
To: [email protected]
Subject: Re: [Jprogramming] xkcd 356
In my opinion, in most cses computer solutions are much esier using nodal
admittance methods once you select a node as a reference then it is very
easy to set up an admittance matrix yii is the sum of admittances connected
to node i from all nodes including the reference.
yij=yji =-sum of admittances between i and j. There is no row or column
related to the reference node.
Invert this matrix to get a Z-Bus matrix which is a generalized Thevenin
impedance The resultant zii terms are the driving point impedances (thevenin
impedance between i and reference.
for impedances between nodes i and j use zii+zjj -2zij which is the thevenin
impedance between i and j Zbus can be built in steps avoiding inversion of a
large matrix.
This method is shown or should be shown in modern power system texts as it
is very useful for fault analysis of large systems.
Don
----- Original Message -----
From: "Aai" < <mailto:[email protected]> [email protected]>
To: <mailto:[email protected]> [email protected]
Sent: Tuesday, January 8, 2013 9:23:43 AM
Subject: Re: [Jprogramming] xkcd 356
In fact this method is like measuring resistance with an okmmeter between
points A and B of the network.
FWIW here is my interpretation/translation of the Maxima code:
nbrs=: 1 3 5 7&{@,
swz=:0,.~0,.0,~0,]
resistornw=: 4 :0
'h w'=.x
'al bl'=. (+w&*)~/"1 y
d =.( <mailto:*=/~@i.@#> *=/~@i.@#) #&> ix=.,3 3 <@(<:#~0~:])@nbrs;. _3 swz
1+i.h,w r=. 1 al }zv=.0$~h*w A=. %. r (al)} d + _1:`[`]}&zv&>ix bl { A +/ .*
1 bl }zv
)
Your example is calculated like:
2 2 resistornw 0 0,:1 1
1
Or as David pointed out:
% +/ % +/"(1) 1 1 ,:1 1
1
On 08-01-13 17:24, Raul Miller wrote:
> I'm a bit confused by the maxima solution. For example, A is both a
> specific node and it's also a 10 by 10 matrix. Similar for B. I can
> assume that the matrix values are something like potential
> contribution from the named node. But... k looks like the 10 10 #.
> of an index pair (with a bit of off-by-one since the maxima solution
> is using 1 based indices) except it's used as an index into A and B.
> And... so maybe A and B are really representing logically different
> kinds of data in different parts of their structure? But what is the
> logical structure then.
>
> ... anyways, the purpose of some of this code is not clear to me.
>
--
Met vriendelijke groet,
@@i = Arie Groeneveld
----------------------------------------------------------------------
For information about J forums see <http://www.jsoftware.com/forums.htm>
http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see <http://www.jsoftware.com/forums.htm>
http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
