designing the different test specimen shapes needed to reach a uniform stress/strain state in the gauge area. With the development of full-field measurement techniques, several novel inverse techniques have been proposed to process the heterogeneous stress/strain fields to identify the whole set of constitutive parameters [7], among which the Virtual Fields Method (VFM) [8]. The Virtual Fields Method enables characterization of the material properties directly from full-field measurements. This method takes advantage of the heterogeneous strain fields obtained through full-field measurement techniques, such as Digital Image Correlation (DIC) [9] in order to obtain more information from a single test. Since the heterogeneity of the strain fields plays an important role in the VFM identification, the accuracy of the identification of the elastic stiffness coefficients depends heavily on the test configuration as well as on the full-field measurement parameters such as camera noise, spatial resolution and smoothing levels. Therefore, the design of the experiment becomes a non-trivial issue when using the VFM technique. A methodology to optimize the test configuration for VFM identification was firstly proposed by Pierron et al. [10]. The idea was to find an optimized specimen length and orthotropic material axis angle so as to minimize a cost function based on the sensitivity to noise of the sought material stiffness components. Recently, a refined test configuration design procedure was proposed by Rossi and Pierron [11]. The study used the grid method as the full-field technique and simulated the whole measurement and identification chain, including image forming and grid method algorithm. This study provided a significant improvement of the optimization procedure by introducing the many different types of error sources into the cost function. In particular, the effect of the spatial resolution of the full-field technique was correctly taken into account which was not the case in [10]. However, this approach was not validated experimentally. Also, it relied on the grid method (also known as sampling moire´) which is not so commonly used in the experimental mechanics community. In a recent article, an efficient experimental methodology to identify all the material stiffness parameters of a PVC foam material in one single test using Digital Image Correlation and the Virtual Fields Method was presented [12]. The study provided an optimized material test configuration with a particular objective of improving the accuracy of the identification. Although the selected test configuration led to a significant improvement of the experimental results, significant differences were found between reference values of material parameters known from literature and/or other tests and the experimental results from this that study. It was thought that one of the reasons for this is that the conducted optimization study was based on finite element simulated strain fields which do not include the sources of error that arise from real DIC measurements. In particular, the low-pass spatial filtering effect of the DIC measurements will lead to underestimation of the strains in large strain gradients areas of the test specimen, which in turn will lead to biases on the identified stiffness components. Moreover, the low signal to noise ratio associated with the measurement of the elastic material properties of polymer foams will tend to increase the random error (scatter) of the data. In the present study, a procedure has been developed to realistically simulate the modified Arcan test for polymer foams using Digital Image Correlation and the Virtual Fields Method. The idea is to construct deformed synthetic images using finite element (FE) displacements. From this, the reference and deformed synthetic images will be processe dusing Digital Image Correlation (DIC), and the Virtual Fields Method 5VFM) will be used subsequently to extract all the stiffness parameters. This paper aims at reporting a few preliminary result of this study. 18.2 Modified Arcan Test In order to generate heterogeneous states of stress and strain in the foam specimen, a modified Aran fixture is used as in [12]. Thick orthotropic Divinycell H100 PVC foam panels from DIAB are considered here. The orthotropy directions are denoted ‘1’ (longitudinal), ‘2’ (through-the-thickness direction) and ‘3’ (depth direction). The specimens can be cut from a thick foam panel such that that the ‘1’ direction lies at a certain angle θ from the specimen longitudinal axis ‘x’ (see Fig. 18.1), resulting in a material ‘off-axis’ configuration. Besides, the modified Arcan fixture enables to load this specimen at a certain angle α from the ‘y’ direction of the specimen (see Figs. 18.1 and 18.2). In [12], a systematic study showed that certain configurations (i.e., certain sets of values for α and θ) produced better identification results than others. However, this analysis was based the strain maps from simulated data and did not allow to derive expected uncertainties reliably. This is the objective of the present work. 138 P. Wang et al.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==