Predicting the strength of concrete made with stone dust and nylon fiber using artificial neural network

Sourav Ray; Mohaiminul Haque; Tanvir Ahmed; Ayesha Ferdous Mita; Md Hadiuzzaman Saikat; Md Mafus Alom

doi:10.1016/j.heliyon.2022.e09129

Predicting the strength of concrete made with stone dust and nylon fiber using artificial neural network

Heliyon. 2022 Mar 18;8(3):e09129. doi: 10.1016/j.heliyon.2022.e09129. eCollection 2022 Mar.

Authors

Sourav Ray¹, Mohaiminul Haque¹, Tanvir Ahmed¹, Ayesha Ferdous Mita¹, Md Hadiuzzaman Saikat¹, Md Mafus Alom¹

Affiliation

¹ Department of Civil and Environmental Engineering, Shahjalal University of Science and Technology, Sylhet, Bangladesh.

Abstract

Excessive demand of concrete is causing depletion of natural sand resources. Especially, the extraction of river sand negatively affects its surrounding environment. A sustainable solution to this problem can be the proper utilization of waste materials and by-products like stone dust (SD) as fine aggregate replacement in concrete. The recycling of stone dust as a construction material lessens the use of natural resources and helps to solve landfill scarcity as well as environmental problems. Addition of nylon fiber (NF) as fiber reinforcement can also attribute to enhance the properties of concrete. This research aims at utilizing SD as fine aggregate along with NF, and assessing the compressive strength and splitting tensile strength of concrete. Although the individual effects of incorporating stone dust and nylon fiber in concrete have been investigated in previous researches, their combined effects, as well as effects of water cement (WC) ratio on concrete strength, have not been studied yet. In this study, volumetric percentages of stone dust (20%-50%) and nylon fiber (0.25%-0.75%) and three different water cement ratio (0.45, 0.50 and 0.55) have been considered as three independent variables to develop probabilistic models for compressive strength and splitting tensile strength of concrete using artificial neural network (ANN). The values of coefficient of determination (R²) and other statistical parameters of the developed probabilistic models indicate the accuracy of the models to predict the concrete strength. In terms of compressive strength at early age, the optimal percentages of SD and NF have been found as 20% and 0.25%, respectively. However, the strength gradually drops as water cement ratio elevates from 0.45 to 0.55. The reduction of the splitting tensile strength has been observed for increasing SD from 20% to 50%, whereas, strength increases for rising NF and WC up to mid-level.

Keywords: Artificial neural network; Compressive strength; Nylon fiber; Splitting tensile strength; Stone dust.