GeomLib_Tool::Parameters failure

Hello,

I'm trying to plug some OCCT functions to my own geometric mesher (controlling choral error) in order to generate a mesh for each surface in the bearing.iges file provided with the OCCT distribution.

What basically my mesher need is :

- to get the parameters (u,v) for a given point on the surface ( I use : GeomLib_Tool::Parameters)

- to get the reciprocal (I use : Geom_Surface::Point)

- to get the local derivatives on a given point on the surface (I use D0,D1,D2 from Geom_Surface).

The main problems I have are with GeomLib_Tool::Parameters that fails sometimes :

1) The first issue I had is with MAXTOLERANCEGEOM in the source file that is hard coded and is very restrictive (too many calls return Standard_False). I have increased its value and it give still good result.

- why the value has been hard coded and set to 1.e-4 ? would it be possible to pass it as a parameter or make the users able to change it dynamically ?

2) But even by increasing the value, I still get errors like the one depicted in the attached images. For the bearing case, the surface orange (fig1 for a global view, fig. 2 for a zoom) is not transformed properly (i.e. in an isomorphic manner) as shown in fig. 3. where 2 transformed points get coincident. Do you know what is the issue, if I'm doing something wrong or if there is a workaround or a better way to get the transformed boundary of a surface.

To give you extra details about how I discretize the surface boundaries :

I use GCPnts_UniformDeflection whith a given chordal error. it gives me a number of new points N1 over the contour.

I then compute the length of each edge using GCPnts_AbscissaPoint::Length. Given a input size h, this gives the number of new points N2.

I finally actually do the discretization using max(N1,N2) with GCPnts_UniformAbscissa.

 

Many thanks for your help

Sâm

Attachments: 
Benjamin Bihler's picture

Hi,

I am not sure whether I have understood your problem correctly. But it might be related to something I have asked not long ago: http://www.opencascade.com/content/mapping-topodsface-space-not-bijective

Regards,

Benjamin

sam landier's picture

Thanks Benjamin,

It is indeed related. I have followed the Mauro's explanations and in my case it is not a periodic or surface of revolution issue, so I will try to ask him directly.

Regards,

Sâm