A review of soft computing techniques in predicting the compressive strength of concrete and the future scope

Structural design of Reinforced Cement Concrete (RCC) highly depends on the compressive strength of the concrete used. The compressive strength determination techniques are categorized as destructive, non-destructive, and partially destructive. In non-destructive techniques, the equipment is costly and needs expertise. The compressive strength of concrete is influenced by multiple parameters and materials used in making the concrete. Soft computing techniques like Machine learning (ML) and artificial intelligence (AI) have been proven to find hidden relations between multiple parameters and achieve the desired result. The inclusion of AI/ML has enabled the characterization of the strength with advanced techniques based on the individual constituents or images using digital image correlation. Based on the literature reviewed in this study, ML and AI techniques have shown promising outcomes in predicting the compressive strength of concrete. This study systematically examines the contributions made to date in predicting compressive strength utilizing AI-ML-based strategies. It compares and highlights existing literature based on the type of machine learning techniques used, datasets used, evaluation parameters, and performance of different methods. The study does not encompass high strain rate loading or dynamic type of loading. This paper also aims to find the gap in the research conducted and state the potential scope of estimating compressive strength using soft computing techniques.