PREDICTION OF GEOPOLYMER CONCRETE COMPRESSIVE STRENGTH UTILIZING ARTIFICIAL NEURAL NETWORK AND NONDESTRUCTIVE TESTING

Abstract

A promising substitute for regular concrete is geopolymer concrete. Engineering mechanical parameters of geopolymer concrete, including compressive strength, are frequently measured in the laboratory or in–situ via experimental destructive tests, which calls for a significant quantity of raw materials, a longer time to prepare the samples, and expensive machinery. Thus, to evaluate compressive strength, non–destructive testing is preferred. Therefore, the objective of this research is to develop an artificial neural network model based on the results of destructive and non–destructive tests to assess the compressive strength of geopolymer concrete without needing further destructive tests. According to the artificial neural network analysis developed in this study, the compressive strength of geopolymer concrete can be predicted rather accurately by combining the results of the non–destructive with R2 of 0.9286.