Radiography-based mechanical identification

Clément Jailin (1) , A Buljac , A. Bouterf (1) , F. Hild (1) , S. Roux (1)

1 LMT - Laboratoire de Mécanique et Technologie

A recently developed Projection-based Digital Image Correlation (P-DVC) method is here extended to 4D (space and time) displacement field measurement and mechanical identification based on a single radiograph per loading step instead of volumes as in standard DVC methods. Two levels of data reductions are exploited, namely, reduction of the data acquisition (and time) by a factor of 1000 and reduction of the solution space by exploiting model reduction techniques. The analysis of a tensile elastoplastic test composed of 127 loading steps performed in 6 minutes is presented. The 4D displacement field as well as the elasto-plastic constitutive law are identified. Keywords-Image-based identification, Model reduction, Fast 4D identification Introduction The identification and validation of increasingly complex mechanical models is a major concern in experimental solid mechanics. The recent development of computed tomography coupled with in-situ mechanical tests and full field measurements offers the opportunity to identify kinematic and/or mechanical quantities on the tested sample. The 4D frameworks [1] based on a series of 3D scans (each composed of approximately 1000 radiographs requiring 1 hour scan time in lab-CT) at different loading steps have proved to be very efficient but extremely time intensive (and in turn fragile with respect to slow drifts). These approaches are thus often limited to few scans and forbid time dependent mechanical characterizations. A recently developed Projection-based Digital Volume Correlation (P-DVC) [2] overcomes these difficulties by extracting the measured quantities directly from few radiographs instead of reconstructed volumes. The technique was originally developed for crack characterization in synchrotron or lab-CT [3] by using only two orthogonal projection angles per step resulting in a gain of 300 in acquisition time compared with the classical full scan method. P-DVC is herein extended to elastoplastic identification from a single projection per loading step. The proposed method allows on-the-fly radiograph acquisitions to be carried out while continuously loading and rotating the sample with an additional improvement of more than 2 in the acquisition time. P-DVC is based on the minimization of the quadratic difference between the projection of the deformed sample acquired at angle θ(t), g(r,t), and the computed projection of the reference 3D image,

Domaines: Mécanique des matériaux [physics.class-ph]; Matériaux et structures en mécanique [physics.class-ph]; Matériaux; Traitement du signal et de l'image [eess.SP]

Clément Jailin, A Buljac, A. Bouterf, F. Hild, S. Roux. Radiography-based mechanical identification. PhotoMechanics, Mar 2018, Toulouse, France. ⟨hal-01960376⟩