Personal tools
You are here: Home Research Mobility Exchange Reports Rosalinda Ferrandes Research Summary
Related terms
Navigation
 
Document Actions

Rosalinda Ferrandes

Mobility Exchange Report: INPG->IMATI (May, 30th - July, 29th 2005)

Research Summary

Physical data transport on the different representations of a shape

The generation of simulation models, starting from the design model, needs various steps of adaption, idealization and detail removal. The quality of FE analysis can be strongly influenced by the shape simplification executed on the component. After a simulation on the simplified part, mechanical criteria can help to evaluate the influence of each removed feature. These criteria can alsobe used in an adaptive process of simplification. The simplified part is redefined by adding some of the removed features according to their mechanical influence . An a posteriori estimator has been developed at the Laboratoire 3S. The aim of my research is to automate the complete adaptive process ,which still needs a great interaction of the user.

Identification and subsequent removal of details is performed, in the context of my research, utilizing an intermediate polyhedral model, made up by triangular faces. An intermediate polyhedral model can be considered an appropriate basis for the FE preparation phase.

One important phase of the work consists in process of simplification, with the detection and the selection of the features to suppress.

The work performed during the two months spent in Genova, has focused on the process of detection and localization of both through and blind holes.

Approach

The developed approach is based on the application of the Euler rule. In fact, applying the generalized Euler rule to a model, I know the number h of topological details (not only holes but also handles) in the object.

I have developed an algorithm, able to divide the initial model in two subparts through a cutting plane. Afterwards, I calculate the number of topological details belonging only to one subpart, as well as topological details crossing the cutting plane.

This is a purely virtual subdivision. If some faces, edges or vertices cross the cutting plane, I virtually split them, assigning one entity to both subparts. I need to count the number of vertices, edges and faces to apply the Euler rule to each subpart. I will need also to calculate the number of boundaries and shells that are created by the cutting plane.

If I go on cutting the model by planes both parallel and orthogonal to the first one, I can further localize topological details.

fig1 fig1
by Odd_Andersen last modified 2006-01-11 15:35
« May 2012 »
Su Mo Tu We Th Fr Sa
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31