Evaluating the effect of nanofillers on cement-based composites strength via artificial neural network and genetic algorithm

Abstract

This paper investigated the contributions of nanofillers’ characteristics and contents on the strength of cement-based composites. For this purpose, an artificial neural network and genetic algorithm model was developed based on 150 compressive/splitting tensile experimental results of cement-based composites, comprehensively considering chemical compositions (silica, titania, zirconia, carbon, and boron nitride nanofillers), physical dimensions (minimum size, aspect ratio, and diameter-thickness ratio), surface treatments (surface hydroxyl treatment and surface nickel coating), and contents of nanofillers. The research results indicated that the chemical compositions, physical dimensions, surface treatments, and contents of nanofillers present distinct effects on the compressive strength and splitting tensile strength of cement-based composites. The chemical compositions and surface treatment of nanofillers are the primary factors determining the compressive strength of nano-engineered cement-based composites, and the influence weights of carbon and surface hydroxyl treatment are 41.6% and 12.0%, respectively., respectively. In contrast, the splitting tensile strength of nano-engineered cement-based composites is notably affected by the diameter-thickness ratio and minimum size of nanofillers, and the corresponding influence weights equal to 28.2% and 17.3%, respectively. These findings provide a guideline for designing and controlling the strength of nano-engineered cement-based composites.