Statistical Overlay of CFRP Strain Fields from Digital Image Correlation for Finite Element Analysis Validation

This work aims to enable the practical use of low-fidelity modelling of composite components under extreme load cases, by quantifying the uncertainty and error, contributing towards a framework enabling such models to be utilised in the design process. A key step towards useful and computationally cheap simulations is to accurately represent the behaviour of Carbon Fibre Reinforced Polymer (CFRP) structures which is challenging due to their heterogeneous and stochastic properties.In this study, full-field strain measurements obtained through Digital Image Correlation (DIC) are compared against a corresponding Finite Element Analysis (FEA) model of D5766 open hole tensile CFRP specimens. The experimental strain fields are overlaid, and statistical analysis is performed, before being mapped onto the finite element model allowing pointwise comparison of strain. Differences between the strain fields are spatially mapped, and the discrepancies quantified, identifying regions where the model and experiment diverged, evaluating model accuracy. A classical lamination theory (CLT) finite element model can capture 67% to 75% of the strain field within one standard deviation in the loading direction, with the error localised around the hole due to the overlaying of multiple optical measurement fields and laminate heterogeneity.