Performance analysis of cost‐effective hybrid polypropylene‐steel engineered cementitious composites and prediction based on artificial neural network

Abstract Engineered Cementitious Composites (ECCs) is a high‐performance fiber‐reinforced cementitious composite renowned for their excellent tensile strain hardening behavior. Among the reinforcing fibers, PolyVinyl Alcohol (PVA) featuring sufficient tensile strength but the fiber's cost is relatively high. Hence, this research has attempted to obtain strain‐hardening ECC by locally available fibers. Hybridization of Recron 3s PolyPropylene (PP) fibers and hooked‐ended steel fibers (SFs) were chosen to improve the strain hardening behavior. According to the compressive test and direct tensile test results, mixes developed with 1.7% PP and 0.3% SF displayed better strain corresponding to ultimate load due to the better dispersion of the fibers. Toughness indices and residual strength factor were also estimated using ASTM C 1018 standards. Based on the observations, slabs contains 1.9% PP along with 0.1% SF has a significant impact on toughness indices and residual strength factor. However, the fracture energy increases by increasing the equivalent steel reinforcing indices from 20 to 55. Finally, the cost of the composites has been compared by frequently used PVA‐ECCs from the literature. It could be concluded that the cost could be reduced by around 75% by using the currently used hybrid ECC rather than PVA‐ECC. This study also presents the prediction of mechanical properties of hybrid ECC based on the artificial neural network. In conclusion, feedforward neural networks using a backpropagation algorithm and two processing layers demonstrated adequate ability to predict the mechanical properties of hybrid ECCs.