Bond behavior of galvanized iron fiber reinforced concrete with recycled coarse aggregate and model prediction using machine learning

Concrete is a brittle material with low tensile strength, requiring reinforcement bars to carry the tensile load and ensure structural serviceability and durability. This study aims to improve the mechanical properties and bond behavior of natural aggregate concrete (NAC) and recycled aggregate concrete (RAC) by incorporating locally available galvanized iron fiber (GIF). Two concrete strengths (30 MPa and 40 MPa) were considered with GIF lengths of 15 mm and diameters of 0.5 mm. Eighteen mix combinations were tested with varying GIF (0 %, 0.25 %, 0.5 %) and recycled coarse aggregate (RCA) contents (0 %, 30 %, 50 %). Three rebar diameters (12 mm, 16 mm, and 20 mm) with embedment lengths of 8D and 12D were used. Results showed significant improvements in compressive strength and split tensile strength, up to 39.3 % and 13.93 %, depending on the GIF and RCA percentages. Up to 40.8 % and 46.5 % higher bond strength was found using 0.25 % and 0.5 % GIF, respectively. The study also employed regression and machine learning (ML) models to predict bond strength. The XGB and ANN models were used to compare the proposed regression equations and existing mechanical models with the ML models. Based on the investigation, it is suggested that 0.25 % or 0.5 % of GIF be used while limiting the RCA content to 30 % for optimal performance. By utilizing locally available and cost-effective GIF alongside RCA, these findings contribute to sustainable construction practices by enhancing the mechanical and bond properties of concrete while addressing environmental concerns.