Forecasting Compressive Strength of Fly-Ash-Based Geopolymer Concrete using Gene Expression Programming (GEP)

Abstract

To produce geopolymer concrete (GPC), fly-ash (FA) like waste material has been effectively utilized by various researchers. The laboratory methods to find the compressive strength (𝑓𝑐′) of fly-ash based geopolymer concrete (FGPC) are expensive, time-consuming, and require skilled personnel. In this research, the soft computing techniques known as gene expression programming (GEP) were executed to deliver an empirical equation to estimate the 𝑓𝑐′ of FGPC. To build a model, a consistent, extensive and reliable data base is compiled through a detailed review of the published research. The compiled data set is comprised of 298 𝑓𝑐′ experimental results. The 10 utmost dominant parameters are counted as explanatory variables, in other words, the extra water added as percent FA (%𝐸𝑊), the percentage of plasticizer (%𝑃), the initial curing temperature (𝑇), the age of the specimen (𝐴), the curing duration (𝑡), the fine aggregate to total aggregate ratio (𝐹𝐴𝐺⁄), the percentage of total aggregate by volume ( %𝐴𝐺), the percent SiO2 solids to water ratio (% 𝑆/𝑊) in sodium silicate (Na2SiO3) solution, the NaOH solution molarity (𝑀), the activator or alkali to FA ratio (𝐴𝐿𝐹𝐴⁄), and the Na2SiO3 to NaOH ratio (𝑁𝑠𝑁𝑜⁄). A GEP empirical equation is proposed to estimate the 𝑓𝑐′ of FGPC. The accuracy, generalization, and prediction capability of the proposed model was evaluated by performing parametric and sensitivity analysis, applying statistical checks, and then compared with non-linear and linear regression equations. The performance index (𝜌) for training set and validation set approaches to zero, witnesses the better GEP model. In the validation stage, the 𝜌 reveals that GEP model is 53% and 46% better than linear and non-linear regression models respectively. The model correctly meets the appropriate requirements for external validation considered. The validation of the proposed GEP model via experimental results shows that it possesses higher generalization and predictive capability and is appropriate to practice in the preliminary design of FGPC according to the Pakistani environment.