An Optimized Statistical Model for Predicting Composite Modulus of Elasticity

Abstract Composites are the revolutionary materials that were developed for the ease of the technology. Similar to all families of materials, composites are being extensively studied nowadays. One of the composites ‘main studies is the homogenization study to determine composites modulus of elasticity function of multiple variables based on differentiating several inclusions’ geometries, or quantities or orientations. However, homogenization studies require extensive numerical and analytical work. This work uses a statistical optimized tool TSREG (tabu search combined with statistical regression proven to achieve models with the highest R-squared and lowest p-values for each variable in addition to the lowest MAPE (mean absolute percentage error)) to predict a model relating composite modulus of elasticity to inclusions shape (as aspect ratio), volume fraction, and orientation (0-30-60-90°). Experimental data of modulus of elasticity compressed 3D printed ABS plastics cubes of 16 × 16 × 16 mm3 size having one inclusion (as empty spheres or ellipsoids with zero Young’s modulus) were utilized. For the voids, their geometries were varied to cover spherical and elliptical shapes with several aspect ratios (0.2-0.4-0.65-0.75-0.9-1), volume fractions (0-0.1-0.2-0.3-0.35-0.4-0.5-1), and orientations (0-30-60-90°). This model helps researchers to determine the composite modulus of elasticity using one significant and accurate expression without using numerical analysis.