Synergy performance of hybrid fiber-reinforced ultra-high-performance cementitious composites with low fiber contents

This study aims to assess the synergistic tensile performance resulting from the hybridization of long and short fibers. Three types of long steel,fibers, i.e., twisted, hooked, and smooth fibers, along with two types of short fibers, i.e., smooth and polyamide fibers, were incorporated into ultra-high-performance concrete (UHPC) at a total volume content of 1.5%. To predict the tensile resistance of the hybridizations, various machine learning models, including artificial neural network (ANN), decision tree (DT), random forest (RF), and support vector machine (SVM), were applied by utilizing a significant number of collected experimental results. Experimental findings demonstrated that the hybridization of long and short fibers effectively enhanced tensile resistance compared to mono fibers. These hybridizations exhibited negative synergy factors in post-cracking strength but positive synergy factors in both strain capacity and specific work to fracture. Predictions using machine learning models revealed that the RF model exhibited outstanding performance in predicting the tensile resistance of the hybridizations. Furthermore, the compressive strength of the matrix was found to be the most important factor affecting post-cracking strength, whereas fiber length had the most substantial impact on the strain capacity.