Synthesis of geopolymer mortar from biomass ashes and forecasting its compressive strength behaviour

Abstract

The global warming dilemma raised an exigency toward sustainability in the construction industry and compelled scientists to hunt for alternative binders in construction materials. The, geopolymerization as a spectacular technology has arisen as a looming solution to this dilemma. This work assesses the features of geopolymer mortar developed with both industrial and agricultural wastes at ambient temperature curing. The geopolymer mix proportion was evolved in this study using main precursor material as fly ash, ground granulated blast furnace slag (GGBFS), sugarcane bagasse ash (SCBA) or rice husk ash (RHA). Workability and mechanical properties, including compressive strength, flexural strength, and split tensile strength of different mortar mixes with various percentages of RHA and SCBA for 365 days, were evaluated. Geopolymer mortar with 10 % SCBA and 5 % RHA achieved improvements of 25 % and 13.91 % in compressive strength, 10.6 % and 3.5 % in flexural strength, and 35 % and 16.8 % in split tensile strength, respectively. Good geopolymer gel formation is clearly evident from the SEM images of SCBA mortar. The main polymeric bonds Si–O–Al and Si-O-Si are also identified using FTIR. To assess the efficacious use of biomass ashes, resistance to chemical attack using H2SO4, HCl, Na2SiO3, and NaCl solutions for 365 days was studied. It was revealed that SCBA incorporated geopolymer mortar specimens achieved good resistance compared with RHA geopolymer mortar. The relation between the ambient temperature cured geopolymer specimen compressive strength at 28 days and 365 days chemical solution curried geopolymer specimen compressive strength find out using python and obtained R2 values more than 95 %. Also, this research proposed various machine learning techniques such as long short-term memory, support vector regression, random forest regression, XGBOOST, and AdaBOOST algorithms for the prediction of compressive strength of geopolymer using 724 mixture proportions with diverse curing temperatures and varying ages. The XGBOOST algorithm achieved the highest R² value of 0.92, demonstrating its effectiveness for the strength prediction.