Enhancing the strength of fly ash (FA)-based geopolymer by increasing the alkaline activator content is a costly and unsustainable technique. Therefore, this work was designed to reduce the activator by employing the pressured catalysis (PC) technique, coupled with the use of minerals that have filler and occupying effects. The main objective was to enhance the strength of the mix with a lower alkaline-to-precursor (A/P) ratio and create a sustainable, load-bearing building block from it. Initially, the compressive strength of the FA-based geopolymer was investigated experimentally by varying sodium silicate to sodium hydroxide and A/P ratios with ambient and hot curing. Afterward, PC was applied to the optimized proportion of constituents, and a significant increase in strength (9.6 to 20.0 Mpa) was observed at a 0.25 A/P ratio. By adding clay and dune sand (DS), the compressive strength was 19.5 and 40.4 Mpa at an A/P of 0.25 and 0.16, respectively. The strength gain mechanism was evaluated at the molecular and micro levels by conducting FTIR and SEM analyses. The environmental and economic indices and strength indicated the high sustainability of DS-based geopolymers compared to analogous blocks. The environmental and economic benefits of 23.9% reduced CO₂ emissions and 24.2% less cost were provided by the DS-based block compared to the FA–clay-based block. A DS-based geopolymer obtains strength at a low A/P due to its occupying effect and results in sustainable building blocks.

The static compaction technique emphasizes the reduced activator dosage required to develop geopolymers. Therefore, it is crucial to comprehend the optimal alkaline activator concentration for blending low-calcium precursor (fly ash) with high-calcium precursor (GGBS) to produce geopolymer blocks. This work was designed to optimize structural blocks’ compressive strength and durability. In experimentation, fly ash (FA) and slag (GGBS) proportions were initially investigated under NaOH solution with varying molarity (8–12) and curing conditions to develop a load-bearing structural block. Subsequently, the durability of the optimized block was evaluated over 56 days through subjection to sulfate and acidic solutions, with efflorescence monitored over the same period. The results reveal that the structural block comprised of 100% FA exhibits the highest compressive strength and lowest bulk density. Conversely, the block incorporating 25% slag that underwent hot curing demonstrates a remarkable 305% strength increase compared to ambient curing. Considering the physico-mechanical performance, the 100% FA block was chosen for durability investigation. The findings indicate a substantial strength loss exceeding 40% after exposure to sulfate and acidic environments over 56 days, coupled with pronounced efflorescence. Catastrophic failure occurs in all cases due to significant strength deterioration. The FTIR spectrum revealed the shifting of the wavenumber to a higher value and verified the depolymerization and leaching of alumina under acidic exposure. However, the developed geopolymer blocks demonstrate superior sustainability and feasibility compared to conventional fired clay bricks and cement-based FA bricks. Despite slightly higher costs, these blocks exhibit greater strength than their counterparts after enduring severe exposures.

The incorporation of fracture mechanics into structural engineering has underscored the importance of fracture energy in understanding crack propagation in quasi-brittle concrete. Nevertheless, conventional laboratory testing remains labor-intensive, time-consuming, and costly. This study applies hybrid machine learning (ML) algorithms, including ANN-GA, SVR-GA, XGB-GA, GBR-GA, and a Stacking Ensemble, to efficiently predict key concrete properties such as initial fracture energy (IFEC), modulus of elasticity (ME), tensile strength (TS), water absorption (WA), and abrasion resistance (AR). Among individual models, IFEC-XGB-GA and IFEC-GBR-GA demonstrated the most stable performance, while the Stacking Ensemble improved overall prediction accuracy and robustness by approximately 5–10 compared with single models. For ME prediction, the ensemble approach reduced bias and variance, although GBR performed better under extreme conditions. TS-ANN-GA exhibited overfitting, whereas TS-Stacking Ensemble achieved the lowest testing errors (MAE0.145, RMSE0.232) and the highest explained variance (≈1.014) and correlation (0.973), confirming the advantages of model aggregation. Moreover, GBR-GA effectively predicted water absorption (R ² 0.998) and classified abrasion resistance with about 84 accuracy, highlighting strong potential for durability evaluation. A graphical user interface (GUI) was developed to integrate these models, providing engineers with a practical, data-driven, and physics-informed tool for accurately predicting the fracture energy of concrete.