Search

International audience; Background: Even though the simulations used to describe the failure of laminates are becoming more and more predictive, complex testing under multiaxial loadings is still required to validate the design of structural parts in a wide range of industrial domains. It is thus essential to assess the actual boundary conditions to allow for an objective comparison between testing and calculations, in particular since the structural tests are complex and often leads to buckling. Therefore, accurate estimation of force and moment fluxes applied to the specimen is critical. In this context, stereo DIC has proven to be an important measurement tool and provides very well resolved surface displacement fields, but the exploitation of such measurements to calculate fluxes remains problematic when testing composites. Objectives: The first objective of this study is both to reduce the uncertainty associated with fluxes determination on a complex test and to simplify the extraction process with respect to existing procedures. The second objective is to make this methodology robust to geometrically nonlinear deformations. Methods: In this paper, we propose a new methodology that extracts minimal boundary conditions in the form of 3D mechanically admissible displacements fields. The approach developed uses a Finite-Element Stereo Digital Image Correlation (FE-SDIC) method regularized by the means of mechanical behavior admissibility equations. Results: Results show that the new methodology outputs much more accurate fluxes than classical data generated from multiple differentiations of the displacement fields. Excellent noise robustness is obtained and quantified. Conclusions: Numerical predictions have been satisfactorily compared with experimental data from one structural scale composite specimen under complex testing.

International audience

International audience

Ever increasing weight, size and power demand limitations, ask for reliable, and modern material models, which in turn rely strongly on advanced identification and validation techniques. One of these techniques is the Stereo-Digital Image Correlation (S-DIC), which is increasingly used in experimental mechanics for full-field kinematic measurement. Recent developments of a global S-DIC formalism, based on finite element, allows to measure a 3D-kinematic field directly in the object coordinate system. A benefit of this method is that the experimental measurement is expressed in the same discretization support as the simulation, which allows to use the model in the measurement and to greatly simplify the dialogue between experiments and numerical simulations.A “learning phase” of the experimental setup precedes the displacement measurement, where the system learns the context of the scene. It is composed of (1) learning about characteristics and placement of the cameras; (2) learning of the true shape of the object by correcting the initial CAD shape. Using S-DIC framework the (1) calibration, and (2) shape measurement steps in learning phase are written in the unified manner. In addition, chosen procedure allows: simplifie initialization of the algorithms with feature marks; improve the boundary management with using of masks; limit reflections influence with an element by element adjustment of brightness. Finally, a noninvasive geometric regularization, based on iso-geometric analysis, is used.The performance of the proposed procedure is tried with a tension test on a stainless-steel sample. After the learning phase, the experimental setup is used for displacement field measurement. Then, the identification of parameters for an elastoplastic law is performed. For the identification procedure, the same finite element mesh is used both for displacement measurement and simulation within a finite element model updating (“FEMU”) algorithm.The presented methodology is implemented in an academic Python program.; Les limitations croissantes de poids, taille et puissance exigent des modèles de matériaux fiables et modernes, qui reposent fortement sur des techniques avancées d'identification et de validation. L'une d'entre-elles est la stéréo-corrélation d'images numériques (S-CIN), de plus en plus utilisée en mécanique expérimentale. Le développement récent d'un formalisme de S-CIN global, basé sur les éléments finis, permet de mesurer un champ cinématique 3D directement dans le repère de l'objet. Un avantage de cette approche est l'expression de la mesure dans le même support de discrétisation que la simulation, ce qui permet d'utiliser le modèle dans la mesure et de fortement simplifier le dialogue entre les essais et les simulations numériques.Une « phase d'apprentissage » de la configuration expérimentale précède la mesure de déplacement, où le système apprend le contexte de la scène. Elle se compose (1) de l'acquisition des caractéristiques et du positionnement des caméras ; (2) de l'apprentissage de la forme réelle de l'objet en corrigeant la forme issue de la CAO. Dans le même formalisme que la S-CIN, les étapes (1) d'étalonnage et (2) de mesure de forme, sont exprimées de façon unifiées. De plus la procédure adoptée permet : de faciliter l'initialisation des algorithmes grâce au marquage d'amers ; d'améliorer la gestion des bords par l'application d'un masque ; de limiter l'influence des reflets par l'ajustement de la luminosité, élément par élément. Enfin, une méthode géométrique et non-invasive de régularisation, basée sur l'analyse iso-géométrique, est mise en place.L'efficacité de cette procédure est testée avec un essai de traction sur une éprouvette d'acier inoxydable. Après la phase d'apprentissage, le montage est utilisé pour mesurer le champ de déplacement. Puis, l'identification des paramètres d'une loi de comportement élasto-plastique est réalisée. Dans cette étape, l'utilisation du même maillage éléments finis, à la fois pour la mesure et la simulation, est exploitée par une méthode de recalage de modèle éléments finis (ou « FEMU » en anglais).La méthodologie présentée est implémentée dans un logiciel académique programmé dans le langage Python.

International audience; Finite element stereo digital image correlation (FE-SDIC) requires a crucial calibration phase in which the initial CAD needs to be updated to fit the actual shape of the specimen. On the one hand, the use of a FE mesh facilitates the coupling of measurements with simulation tools. On the other hand, it provides a unique, fine description of both the geometry and the displacement, which often makes the shape measurement problem highly ill-posed. As a remedy, we propose a hybrid isogeometric-FE strategy that can measure a shape in terms of spline functions while considering as an input and output the analysis-suitable FE mesh. Making use of the appealing spline refinement procedures and of Bézier-based operators, multilevel smooth spline dis-cretizations are built concurrently with the initial FE subspace and related to the multi-scale images used for the initialization of the shape measurement. It results in a geometrically sound regularization which provides a spline parametrization of the optimal shape along with its FE twin. A non-invasive implementation from an existing FE-SDIC code is also detailed. The performance of the proposed method is assessed on real images and comparisons are made with other published techniques to prove its efficiency.

International audience; The recent development of a global stereo digital image correlation formalism, based on a finite element mesh, allows to measure a 3D-kinematic field directly in the world coordinate system. A benefit of this method is that the experimental measurement is expressed in the same discretization support as the simulation. Then dialogue between tests and simulations is simplified. "FEMU"-like algorithm can be used for the identification of the parameters of a constitutive law. This "FEMU" approach is applied to the identification of the elastoplastic law of an open-hole steel plate. The results are compared to those obtained from 2D-DIC.; La récente mise au point d'un formalisme global de stéréo-corrélation, basée sur un maillage éléments finis permet de réaliser une mesure cinématique en 3D directement dans le repère physique. Un avantage de cette méthode est l'expression de la mesure dans le même support de discrétisation que la simulation, ce qui permet de simplifier les dialogues entre l'essai et les simulations. Cette caractéristique peut-être mise à profit par un algorithme de type « FEMU » afin d'identifier les paramètres d'une loi de comportement mécanique. Cette méthode « FEMU » est appliquée à l'identification d'une loi élasto-plastique d'une éprouvette trouée en acier. Ces résultats sont comparés à ceux issus d'une identification à partir de corrélation en 2D.