In recent years, Geopolymer concrete (GPC) stands as a promising alternative to traditional Portland cement-based concrete in the quest for eco-friendly and sustainable construction materials. However, the practical application of GPC remains limited due to the need for controlled high-temperature curing environments. Furthermore, accurately predicting GPC properties has become an urgent issue to reduce the time and cost associated with laboratory experiments. This study aims to investigate how different curing methods, for both on-site and precast application, impact the fracture energy of GPC. Three curing methods at varying temperatures, room temperature (25 °C), mobile dryer (50 °C), and heating cabinet (80 °C), were considered. A comprehensive dataset comprising 194 fracture test results was compiled to propose diverse machine learning models capable of forecasting GPC fracture energy, both without and with various fibers (Steel, Polypropylene, Basalt, Polyvinyl alcohol…). The findings revealed that GPC fracture energy increased by 89% when the curing temperature rose from 25 to 50 °C. A further increase in temperature from 50 to 80 °C resulted in an additional 11% increase in fracture energy. Interestingly, GPC exhibited slightly higher (up to 8%) fracture energy than Portland cement concrete when their compressive strengths were comparable. The developed machine learning models exhibited correlation coefficient values above 0.95 and root mean squared error values within 10% of the average actual values, indicating their exceptional accuracy in predicting GPC fracture energy, regardless of fiber presence. Additionally, sensitivity analysis highlighted the significance of fiber volume content as the primary factor influencing fracture energy in GPC, while in GPC without fibers, the notch depth ratio emerged as the determining factor, accounting for 78% predictive importance.
Read full abstract