Axioms (Jan 2023)

Prediction of Mechanical Properties of Rubberized Concrete Incorporating Fly Ash and Nano Silica by Artificial Neural Network Technique

  • Musa Adamu,
  • Andaç Batur Çolak,
  • Yasser E. Ibrahim,
  • Sadi I. Haruna,
  • Mukhtar Fatihu Hamza

DOI
https://doi.org/10.3390/axioms12010081
Journal volume & issue
Vol. 12, no. 1
p. 81

Abstract

Read online

The use of enormous amounts of material is required for production. Due to the current emphasis on the environment and sustainability of materials, waste products and by-products, including silica fume and fly ash (FA), are incorporated into concrete as a substitute partially for cement. Additionally, concrete fine aggregate has indeed been largely replaced by waste materials like crumb rubber (CR), thus it reduces the mechanical properties but improved some other properties of the concrete. To decrease the detrimental effects of the CR, concrete is therefore enhanced with nanomaterials such nano silica (NS). The concrete mechanical properties are essential for the designing and constRuction of concrete structures. Concrete with several variables can have its mechanical characteristics predicted by an artificial neural network (ANN) technique. Using ANN approaches, this paper predict the mechanical characteristics of concrete constructed with FA as a partial substitute for cement, CR as a partial replacement for fine aggregate, and NS as an addition. Using an artificial neural network (ANN) technique, the mechanical characteristics investigated comprise splitting tensile strength (Fs), compressive strength (Fc), modulus of elasticity (Ec) and flexural strength (Ff). The ANN model was used to train and test the dataset obtained from the experimental program. Fc, Fs, Ff and Ec were predicted from added admixtures such as CR, NS, FA and curing age (P). The modelling result indicated that ANN predicted the strength with high accuracy. The proportional deviation mean (MoD) values calculated for Fc, Fs, Ff and Ec values were −0.28%, 0.14%, 0.87% and 1.17%, respectively, which are closed to zero line. The resulting ANN model’s mean square error (MSE) values and coefficient of determination (R2) are 6.45 × 10−2 and 0.99496, respectively.

Keywords