Modeling the impact of supplementary cementitious materials on compressive strength of recycled aggregate concrete forest-random approach

Abstract

Recycled concrete aggregates (RCAs) and supplementary cementitious materials (SCMs) may substitute some cement and natural aggregates (NA) in concrete manufacturing. However, their effects on recycled aggregate concrete (RAC) compressive strength are difficult to model. Reactivity, silica, and alumina modulus were examined for cementitious materials' chemical complexity. Random Forest approaches were developed to predict and analyze RAC compressive strength. Even with RCAs and SCMs, the RF model accurately estimated concrete compressive strength. The Variable Importance (VI) research examined how input factors affected RAC compressive strength. VI indicated that silica fume contributes most to RAC compressive strength, followed by cementitious materials' reactivity modulus, cement content, silica modulus, fine natural aggregate content, and coarse natural aggregate dosage. The water dosage, water/binder ratio, and RCA content lower the RAC compressive strength. As a result, to highlight, the amount of SCM was not significant, but its nature was (i.e., hydraulic, silica pozzolanic, or alumina pozzolanic).

Keywords

• Random Forest algorithm; compressive strength; supplementary cementitious materials; recycled concrete aggregate; reactivity modulus; silica modulus; alumina modulus; sustainability