14:00
Solid Mechanics 1
Chair: Mike Sutton
14:00
20 mins
|
OPTIMIZATION AND VALIDATION OF A CRUCIFORM GEOMETRY FOR THE IDENTIFICATION OF ELASTOPLASTIC MODELS
Morgan Bertin, François Hild, Stéphane Roux, Florent Mathieu, Hugo Leclerc
Abstract: A method based upon the sensitivity fields is proposed to optimize
biaxial experimental procedures. The objective is to improve the
identifiability of elastoplastic parameters related to an a priori
chosen constitutive law. This is achieved by minimizing the
covariance matrix of the material parameters. This approach
unifies full field measurements via image correlation, numerical
simulations and identification procedure.
|
14:20
20 mins
|
A GLOBAL FORMULATION TO STEREO DIGITAL IMAGE CORRELATION WITH MECHANICAL REGULARISATION
Jean-Emmanuel Pierré, Jean-Charles Passieux, Jean-Noël Périé, Florian Bugarin, Laurent Robert
Abstract: The aim of this study is to measure, with stereo digital image correlation, the displacements and rotations at the boundary of a plate subjected to complex loadings. With classical DIC techniques, the measured field is the displacement of the upper skin only. The rotations are usually estimated a posteriori by numerical differentiation and smoothing. Unfortunately additional uncertainties may arise with these steps. Conversely, the use of a finite element model is proposed here to regularise a stereo FE-DIC measurement. For that a formulation using the 3D displacement as primary unknown is introduced. It requires projection operators based on a pinhole camera model. It is shown that it is possible to take into account distortions directly. Associated to this formulation a more accurate quadrature technique is used avoiding a systematic bias.
|
14:40
20 mins
|
“FREEFORM” ELASTO-PLASTIC MODEL IDENTIFICATION WITH MULTI-SCALE INTEGRATED DIGITAL IMAGE CORRELATION
Florent Mathieu, Jan Neggers, François Hild, Stéphane Roux
Abstract: The identification of elasto-plastic parameters has been performed for different constitutive models. In most of these models, some of the parameters are bound together by the model equation, and thus interdependent. This leads to difficulties to identify properly the contribution of each parameter. Moreover, some of the plasticity models are often expressed in the form of tabulated data when used in FE codes. In this context, defining the elasto-plastic model in terms of parameterization of the stress-strain curve can increase the stability of the analysis, as this parameterization is described with chosen shape functions that are easier to be made less dependent. This approach has been tested on a piecewise linear stress-strain curve, using a multi-scale mesh coarsening approach and Integrated Digital Image Correlation (I-DIC). Handling the number of parameters describing the law may ultimately yield ways to evaluate a part of the model error in the identification procedure.
|
15:00
20 mins
|
MODEL ORDER REDUCTION FOR IDENTIFICATION PURPOSES
Jan Neggers, Florent Mathieu, Stéphane Roux, François Hild
Abstract: The sensitivity fields required for identification methods such as Finite Element Method Updating (FEMU) or Integrated Digital Image Correlation (IDIC) are usually expensive to compute and require considerable amounts of memory storage. This research evaluates the application of Proper Orthogonal Decomposition (POD) to provide efficient means to reduce the weight to express these fields. It is shown that, as little as four orthogonal modes are required to adequately express the sensitivity fields needed to identify the parameters of a Voce isotropic hardening model.
|
15:20
20 mins
|
IDENTIFICATION OF HETEROGENEOUS ELASTOPLASTIC BEHAVIORS USING DIC MEASUREMENT
Tarik Madani, Yann Monerie, Stephane Pagano, Céline Pelissou, Bertrand Wattrisse
Abstract: The use of full-field measurements in the identification of material properties is currently widespread thanks to advances in measurements techniques and computer-assisted identification methods. In this paper, an iterative procedure is used to identify the local stress fields and the material properties distributions using full-field measurement techniques. After summarizing the principle of the method, we focus on its validation in which we identify an elastoplastic behavior. Then the method was applied on noisy measured displacement fields to assess its robustness.
|
|