Composite material properties characterization using digital image correlation and finite element model updating: Uncertainty analysis

This work presents an uncertainty analysis of displacement measurement, strain assessment due to intrinsic measurement error in digital image correlation (DIC) technique and reconstruction approximation approach. The effect of such uncertainty on the extracted material properties has been illustrated by a modified short beam shear (SBS) test. The noisy and discrete displacement field has been measured by DIC, and displacement and strain fields are constructed using a global finite-element based approximation method. Using a linear least-squares regression method and finite-element-model updating (FEMU), multiple composite material constants were extracted from the FEM-calculated stress and DIC-reconstructed strain field simultaneously. A key observation was the presence of Gaussian distributed measurement error and approximation reconstruction error in reconstructing displacement field and strain field, which must be tracked along the identification procedure to provide uncertainties for the extracted material constants. The effects of the measurement error in DIC and reconstruction parameter on the uncertainty of the extracted material properties have been investigated. The identified results from the area of interest and the full-field area have been compared. The effect of the number of images on the identified material constants has been discussed. The results demonstrate that the standard deviation of the material constant depends on the size of the displacement field and the number of images used in the identification procedure. It increases linearly with the standard deviation of the Gaussian distributed measurement error and decreases nonlinearly with increasing the reconstruction mesh size. Reconstruction mesh size is a crucial parameter and no material constants can be extracted from the identification procedure reliably as it deviates from an optimal value significantly.