PERFORMANCE EVALUATION AND STRENGTH PREDICTION OF STEEL SLAG-MODIFIED CONCRETE USING MACHINE LEARNING

The increasing demand for natural aggregates in concrete production has sparked significant concerns about resource depletion, environmental sustainability, and rising costs. This study explores the potential of steel slag, a by-product of steel manufacturing, as a sustainable alternative to natural coarse aggregates in concrete. The aim is to assess the performance of steel slag as a partial replacement for coarse aggregates, evaluating its impact on the physical and mechanical properties of concrete. An extensive experimental program was carried out, involving the preparation of concrete samples with varying replacement levels of coarse aggregate with steel slag at 15%, 30%, 45%, and 60% by volume. Standard tests, including slump tests for workability, compressive strength tests, tensile strength tests, and flexural strength tests, were conducted in compliance with ASTM and BS standards. Additionally, a predictive model using the Random Forest (RF) algorithm was developed to estimate the compressive strength of steel slag aggregate (SSA) concrete across different replacement levels. The findings revealled that while increasing steel slag content reduces workability evidenced by lower slump values it significantly enhances compressive, tensile, and flexural strengths. Optimal performance was observed at a 45% replacement level, beyond which strength properties began to decline. This improvement is attributed to the dense and durable nature of steel slag, which strengthens the interfacial bonding within the concrete matrix. The Random Forest model demonstrated exceptional predictive accuracy, achieving an R² value of 0.85, underscoring its reliability in estimating the compressive strength of SSA concrete. This study concludes that steel slag is a viable and sustainable alternative to natural coarse aggregates in concrete production. Its use could offer advantages such as reducing the demand for traditional aggregate and solving the problem of waste generation as well as ensuring structural reliability and performance. Looking ahead, future research should focus on optimizing steel slag replacement across different concrete grades to further enhance its sustainability in construction. Incorporating additional data from various mix ratios and curing conditions into the RF model will improve its accuracy and reliability for practical structural design applications.