[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Batch processing / interfacing



Hi Dave and femm users.!
I find the program fantastic. Since I downloaded the program for around
year ago I almost quitted Vector fields opera 2D. Femm It is much more
simple and fast to explore ideas with. Further you are not limited by the
expensive dongle and don't need tons of Mb free HD space to the
installation. I really like it because of the simple elegant way it is
constructed. 

But femm can and will hopefully be improved over time. The unlucky thing
about it all is the fact that Dave had all the experience and knowledge to
change it. Even so the source code seams very nice it would take others a
bottleneck of time to make improvements. But if it is possible to take
control over femm with another program femm will definitely be comparable,
if not better to use than expensive commercial programs. I also believe
that most of the users are applying femm and don't concern too much about
the solver and theory behind the code. (I have not seen any femm result
there where far away from what practical measurement and opera 2D gave me)

I don't think there is many there really needs some advanced time-stepping,
3D or some extensive core-loss models so the interfacing from other program
is from my point of view the first priority. 

In fact I tried for some time ago to interface it to another program. I had
parametric model of an electrical machine which I could control in Delphi.
When the parameters had been set I used a procedure/function in Delphi
there made an DXF-file. I thought I easily could interface it to femm by
sending key commands to femm. I started by sending commands like 
sendkeysTo(?Ikke-navgivet - femme', SK_ALT_DN + ´F' + ´O´ + SK_A_LTUP);
sendkeysTo(?test2.fem', SK_ENTER);

But I fast realized that there was no hot key for "import dxf" and that it
was also quite difficult to get proper control over the mouse, which is
necessary in the way femm is currently. So I did not made the interfacing. 

If I shall give a suggestion to how the interfacing is done most
appropriate I will believe that a simple command line like in opera and
Autocad would be most useful for the users and probably not a too heavy job
to do. You need to have to have similar text commands to what you do with
the mouse. By doing so you will avoid thinking about fancy linking like
DDE, OLE or active X or what ever there is popular in these day. Text files
is also easy to create in any programming environments or Matlab and it may
thus be very useful for all users just knowing a little of programming.

DLL functions will be extreme nice but I think it will require lots of
documentation before it will be useful. 

I don't think it will be a good idea to tie femm into a program like
Scilab. This will require that the user have to use a program which isn't
the users favourite. Actually I have some experience with Femlab to Matlab.
It is slow and requires lots of memory but it is the only program that I am
in position to use for some special mechanical Eigen-frequency analysis. If
femm in Scilab will work in that fashion it will not be a plus !!!!

Thanks for a wonderful Program 
Peter from Denmark !



Work :
Assistant Prof. , M.Sc.E.E. Email: por@xxxxxxxxxx
Peter Omand Rasmussen WWW: http://www.iet.auc.dk/~por/
Pontoppidanstaede 101, off. 41 Phone-Secretary: +45 96358080
DK-9220 Aalborg East Phone-Direct: +45 96359244
Denmark Fax: +45 98151411