Rational Experimental-Computational Correlation (RECC) of Four-Point Bending Testing of Multilayer Carbon Fiber Composite

• A novel Rational Experimental-Computational Correlation (RECC) approach is developed for predicting composite behavior accurately and refining FE models. • Sample stiffness decays during loading indicating accurate prediction of material damage evolution. • Full-field DIC is showing to be effective in determining local longitudinal strain and effective bending modulus. • Displacement-based rational correlation is recommended as it delivers more reliable results than strain-based correlation in the evaluation of material parameters. • The RECC approach relates the macroscopic component response under four-point bending with the fiber-level microstructural material properties. This study presents a robust Rational Experimental-Computational Correlation (RECC) methodology applied to deformation maps for assessing material behavior and evaluating the optimized elastic modulus. Four-point bending tests were performed on multilayer carbon/epoxy composite samples, and surface deformation was captured using 2D Digital Image Correlation (DIC) to obtain high-resolution displacement and strain fields. These experimental results were compared with finite element (FE) simulation predictions. The effective elastic bending modulus (E eff ) was determined by fitting the simulation to the experimental data using the coefficient of determination (R 2 ) and Pearson correlation coefficient (r) to quantify the accuracy of the model. Key results indicate a decline in the local longitudinal elasticity modulus (E 11 ) during monotonic loading prior to load drop, signaling microscale material damage. This supports the concept of hierarchical material structures, where meso-level properties such as E 11 influence the global mechanical response and damage of macro-level composite structures. Additionally, comparing strain-based and displacement-based evaluations of E 11 revealed similar trends but shifted values. This is attributed to underestimation in strain-based methods caused by excessive smoothing during numerical differentiation. Therefore, a displacement-based approach is recommended. The results confirm strong agreement with experimental observations and validate the RECC methodology as a reliable tool for understanding and optimizing composite performance.