Structured mesh and edge

[Didier Roissé]

Dear all,
I follow a remark of Enrique about Structured mesh and edge :

Le 06/10/2010 22:23, Enrique Escolano a écrit :

> I have a question/comment:
>
> Following the 1.1.0 version of AMELETHDF specification it say for structured
> meshes that the elements
> are defined by six numbers
> imin jmin kmin imax jmax kmax
>
> I was expecting imin<=imax jmin<=jmax and kmin<=kmax, but it seems that this
> is not true in the NTC1-1_FDTD_v2 example for 'edge' elements
> (it's true for 'face' elements, and in the past samples for 'volume' elements)
> e.g. the fists edge is
> 291 120 529 291 120 528
>
> with 529>528 !!
>
> and for faces it could be an additional group 'normal' that define if the sense
> is + or - (the x,y,z is reduntant, it's known by the face element connectivities)
>
> but for edges not exists a similar data of 'normal' (or better said 'tangent')
> to set the direction in the + or - axe direction
>
> I tink that it is a little inconsistent and could use for both cases (face and
> edge) the same style to set the orientation: by the order of i0 j0 k0 i1 j1 k1
> or by an additional 'orientation' +/- field
>
What do you think about?

[cyril giraudon]

Hello,

Is it a semantic problem or a functional problem ?

If I rewrite structured edge definition in Amelet-HDF as follows :

> meshes that the elements
> are defined by six numbers

> i1 j1 k1 i2 j2 k2

There is no more surprise, i1 can be greater than i2.
An p1p2 is the direction of the edge.

I m not sure a simple rule like this one can be written for face
normal.

Best Regards,

Cyril.