Thermo-mechanical performance and machine learning based prediction of residual strength in laterite GGBS geopolymer composites under cyclic thermal loading

Geopolymers are being investigated as a potential alternative to traditional cementitious materials due to the increasing demand for sustainable and fire-resistant building materials. However, few studies have been carried out on their behavior under cyclic thermal loading. Cyclic thermal loading provides a better simulation of realistic fire exposure conditions and gives a better assessment of the thermo-mechanical performance of the geopolymer ceramics. The present study utilizes a combination of laterite and GGBS as the main precursors for the synthesis of geopolymer. The effect of NaOH molarity (10 M-14 M) and laterite calcination temperature (600–800 °C) on residual compressive strength of geopolymer ceramics after 10 heating-cooling cycles at 200–800 °C is also studied in this research. Results showed that the geopolymer based on 700 °C calcined laterite and activated with 14 M NaOH presented better thermo-mechanical stability in comparison to the other compositions. Furthermore, the microstructural analyses confirmed increased matrix densification, the formation of aluminosilicate gel phases, and a reduction of crack propagation after thermal exposure. Finally, the machine learning algorithms, i.e., Support Vector Regression (SVR) and Multiple Linear Regression (MLR), were built to predict the residual compressive strength under various thermal conditions. The models demonstrated excellent predictive performance, with R² values beyond 0.98. The novelty of this work lies in the integration of experimental investigation and machine learning-based prediction for evaluating the thermo-mechanical behavior of laterite & GGBS geopolymer ceramics under cyclic thermal loading conditions. The findings demonstrate that laterite & GGBS geopolymer ceramics possess significant potential as sustainable fire-resistant materials for high-temperature structural applications, such as tunnels, industrial linings, and thermally exposed infrastructure systems.