Title:
Neural Network for Mixture Design Optimization of Geopolymer Concrete
Author(s):
Ankur Bhogayata, Sneh Kakadiya, and Rinkesh Makwana
Publication:
Materials Journal
Volume:
118
Issue:
4
Appears on pages(s):
91-96
Keywords:
artificial neural network (ANN); compressive strength; geopolymer concrete; optimization; splitting tensile strength
DOI:
10.14359/51732711
Date:
7/1/2021
Abstract:
The paper discusses the development and application of the artificial neural network (ANN) model for predicting the compressive and splitting tensile strength of the geopolymer-based concrete composites (GPC). The strength properties of GPC are influenced by the proportions of the constituents—namely, the alkaline solution, fly ash, aggregates, and sand and water—and require optimization for the desired quality of the composite. The optimum mixture may be obtained by using modern techniques; namely ANN modeling. The ANN models have been developed by training and validating the input data using the sigmoid function and the feed-forward backpropagation algorithm in the hidden layers. The ANN layer is the functional part of the model consisting of the operators to carry out the specific task largely based on mathematical calculations. A five-layered ANN model has been developed and used to predict the strength to optimize the mixture design. The predicted values have been compared with the experimental strength values, and the effects of the most significant constituents have been studied.