Multi-objective optimization and prediction of strength along with durability in acid-resistant self-compacting alkali-activated concrete

Abstract

Sustainability in concrete production reduces the environmental impact and conserves resources, yielding durable concrete. This study focuses on developing and analyzing an acid-resistant Self-Compacting alkali-activated concrete (SCAC). Optimization of mix proportions of SCAC containing GGBFS as a binder was performed through experimental methods and multi-objective optimization. The fresh and mechanical characteristics of SCAC were evaluated, examining the impact of sodium hydroxide (NaOH) molarity and sodium silicate to sodium hydroxide (SS/SH) ratio on strength loss under acid exposure. Optimal mix proportions were identified experimentally for both M25 and M45 SCAC grades. A comprehensive database of 120 observations from experimental findings was used to develop a machine learning (ML) model to predict compressive strength and loss of strength due to acid attack. The gradient boost regression with R2 of 0.96 and the long short-term memory (LSTM) model with R2 of 0.93 were established to be best performing in predicting strength and strength loss respectively. These models were then used as objective functions in a multi-objective optimization to maximize strength and minimize acid-induced strength loss. A cost-based comparison of experimentally and analytically optimized mix proportions was provided. The established optimal NaOH molarity and SS/SH ratio for both M25 and M45 grade SCGC were 12 M and 2.3 along with 12 M and 2.5 respectively. This research offers valuable insights for further experimental work on sustainable and resilient SCAC.