An in-depth exploration of machine learning models for concrete compressive strength and reinforced concrete beam behaviour: A comprehensive review

The mechanical characteristics of concrete, encompassing essential factors such as Compressive Strength, Deflection, Bond Strength, and Shear Strength, are of paramount significance in the design of structurally sound Reinforced Concrete elements. Notably, the assessment of beam deflection is particularly critical for serviceability concerns; however, its accurate determination often requires rigorous analysis or destructive testing. Predicting these vital parameters based on the available test data holds immense value for the design community. Traditional empirical and statistical models, including linear and non-linear regression, have long been employed for this purpose. However, they can be time consuming and prone to errors owing to the variability in factors such as reinforcement, concrete properties, design mix ratios, and curing conditions. To overcome these complexities, researchers have proposed the utilization of Machine Learning algorithms, such as Artificial Neural Networks, Support Vector Machines, Decision Trees, Random Forest, for the prediction of mechanical properties in Reinforced Concrete Beams. The input datasets employed to train these models encompass a range of critical variables, including cement, mineral admixture, Coarse Aggregate, Water/Cement Ratio, Fine Aggregate, Superplasticizer, Slump Value, concrete age, Reinforcement, Reinforcement spacing, Reinforcement, Reinforcement Tensile strength, number of Reinforcement, and Concrete Cover. This study conducts a comprehensive assessment of the applicability and performance of each model, yielding practical recommendations, identifying current knowledge gaps, and delineating potential directions for future research endeavors in this domain.