Abstract
Optimizing the performance of alkali-activated slag backfill grout (ASBG) is crucial for engineering applications. Fluidity and 28 d unconfined compressive strength (28 d UCS) are the two most important performances and multi-objective optimization is existed. In this research, the response surface method (RSM) for experiment design and model regression, non-dominated sorting genetic algorithm II (NSGA-II) for optimization, and entropy-based Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) for decision-making were implemented. The strength formation mechanism of the optimal ASBG was investigated via microstructure evolution. The results show that the quadratic RSM models for fluidity and 28 d UCS exhibit excellent predictive fitting capabilities, with the determination coefficients of 0.9713 and 0.9588 respectively. NSGA-II coupled with entropy-based TOPSIS determined the optimal mix proportion with water-binder ratio of 0.877, binder-sand ratio of 1.000 and alkali dosage of 4.138%. The porosities at 1–28 d decrease from 10.06% to 2.62%. And the mode of microstructure evolution can be divided into two stages according to the probability distribution index and fractal dimension. In early stage, probability distribution index and fractal dimension decreased from 0.9974 to 0.4833 and 1.2498 to 1.2022. This indicated that the number of the small pores decreased and the pore boundary transformed into the softer one as the irregular plate-like GGBS dissolved. In the later stage, probability distribution index and fractal dimension increased from 0.4833 to 1.4732 and 1.2022 to 1.3273, indicating a relatively denser microstructure with fewer large pores formed.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call