Prediction Of Strength Properties Of Geopolymer Concrete Using Artificial Intelligence Techniques

Abstract

Several studies have successfully used fly-ash (FA)-like waste material for the manufacturing of geopolymer concrete (GPC). This study uses gene expression programming (GEP), a type of soft computing approach, to produce an empirical equation that estimates the compressive strength fc0 of GPC using FA. Through a thorough analysis of the published research, a consistent, large, and trustworthy data set is assembled in order to develop a model. 298 fc0 experimental outcomes make up the collected data set. The following are considered explanatory variables: the amount of extra water added as percent FA (%EW), the percentage of plasticizer (%P), the initial curing temperature (T), the specimen's age (A), the curing duration (t), the ratio of fine aggregate to total aggregate (F/AG), the percentage of total aggregate by volume (%AG), the molarity of the NaOH solution, the activator or alkali to FA ratio (AL/FA), the ratio of sodium oxide (Na2O) to water (N/W) for preparing Na2SiO3 solution, and the ratio of Na2SiO3 to NaOH (Ns/No). An empirical GEP equation is put forth to calculate the fc0 of GPC using FA. The suggested model's precision, applicability, and forecasting capacity were assessed using parametric analysis, statistical verification, and a comparison with both linear and non-linear regression equations.