Integrating Rice Husk Ash- Material Science and Artificial Intelligence for Sustainable Cementitious Material Design

Abstract

The increasing demand for sustainable and low-carbon construction materials has intensified research into supplementary cementitious materials derived from agricultural residues. Rice husk ash (RHA), owing to its high amorphous silica content, pozzolanic activity, porous morphology, and potential to reduce Portland cement consumption, represents a promising alternative for developing sustainable cementitious composites. Nevertheless, the engineering performance of RHA-based mixtures is governed by multiple interacting factors, including ash composition and reactivity, burning conditions, particle size, replacement ratio, water-to-binder ratio, curing regime, mixture proportions, and environmental exposure. The nonlinear interactions among these variables make conventional experimental optimization costly, labor-intensive, and difficult to generalize. Demand for low-carbon materials has increased interest in Rice Husk Ash (RHA) as SCM due to its high amorphous silica, pozzolanic activity and cement-saving potential. However, RHA performance depends on many interacting factors - burning conditions, particle size, replacement %, w/b, curing and exposure - making trial-and-error optimization costly.AI/ML [ANN][SVR][RF][GPR] has been used mainly for compressive strength prediction, but durability, service-life and sustainability applications remain limited due to incomplete characterization, heterogeneous data, poor validation and lack of interpretability.[XGBoost]This review links RHA material science with AI, focusing on input selection, feature engineering, explainable, uncertainty-aware and physics-informed ML, and proposes an integrated framework: RHA characterization → predictive modeling → validation → durability/service-life → LCA-based multi-objective optimization, to move from trial-and-error to intelligent design of low-carbon RHA concrete.