Discussion Forum

Hi Kai

Well I like "s" as it's well defined. But s1 and s2 for 2D boundaries are more tricky to use, as they are not necessarily normalised to 0-1 and far less obvious (I find) to understand where they are pointing to.

in V4 the boundary coordinates do help, as also the new vector coordinate plotting options

But on domains, you are left with the Cartesian, and a few common user coordinates systems such as cylindrical, spherical or general mapped, still haven't fully figured out how to use them in equations as they do not have a clear, written out "equation subnode" so some hypothesis and tests are required before using the

...sys2.T11 ...sys2.T12 etc

What else are you thinking of ?

--
Good luck
Ivar

HI Ivar,

until now, I always thought that 0<=(s1, s2)<=0.5

When I used them, the boundary was a rectangle and this usally worked according to the relation above. Now I'm a more carefull using them. :)

I tried 'sys2.T11' and 'sys2.T12' but they weren't defined and I couldn't find something about thm in the doc. I'm (still) using V3.5a, so in V4.1 many things a different ...

I just want to have the equivalent variables of s1 and s2 when using them as boundary parameters at 3D to some variables on the 2D subdomain.


kind regards

Hi

indeed sys2.T11 are part of the transformation matrix of a "user coordinate"system that one can define in V4. The only thing about using them I haven't found the "true" content described in equations, only in "principle", and I'm slightly susceptible, I like formulas that are precise and univoque (thats perhaps a French word, not sure what is the corresponding English / American one ;)

Now I have never trusted that s1, s2 were uniquely normalised, at least there are several warnings about it in the 3.5 doc

In V4 we have the default boundary coordinate system t1, t2, n (in 3.5a too if I remember right) see below

--
Good luck
Ivar

Thanks, Ivar!

I guess switching to 4.x is really needed :)