Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

This paper investigates a human-centered approach for the exploration of large data sets resulting from CFD (Computational Fluid Dynamics) simulations. In a VR (Virtual Reality) immersive environment we propose to couple haptic feedbacks with visual rendering. The proposed solution aims at identifying critical points of an unsteady fluid flow without resorting to any topology analysis such as eigen analysis of the vector field Jacobian. In a first step, while exploring the data volume, a vibration alerts the user on the presence of critical points in the local volume surrounding the probe position. In a second step, critical point properties are emphasized by means of a visuo-haptic feedback. A set of experiments confirms the effectiveness of the proposed approach. Thanks to the combination of haptic feedbacks with the visual materialization of critical points and streamlines, users were able to easily localize, analyze and focus on specific critical points, selected by experts as representative of the flow topology.