Experiment Study on the Effect of Aluminum Sulfate-Based Alkali-Free Accelerator and the w/c on Cement Hydration and Leaching

Abstract

The alkali-free accelerator based on aluminum sulfate is widely used in shotcrete in tunnels. Long-term Ca-leaching of shotcrete may adversely affect the tunnels in a water-rich mountain. It is necessary to examine further the impact of the AS accelerator and w/c on cement hydration and leaching. In this study, all the cement pastes were cured in the environment with R.H. > 95% and 20 ± 1 °C for 60 days and leached in a running water test with 6 M NH4Cl at 1 cm/s. The hydration kinetics was characterized by isothermal calorimetry. Additionally, the microstructural and mineralogical alterations were characterized by XRD, SEM, MIP, and N2 absorption. The results show that (1) the AS accelerator affected the hydration kinetics of cement by stimulating early hydration and delaying the late silicate hydration, resulting in AS-accelerated cement pastes with rougher pore structure. As a result, the higher the dose of AS accelerator, the faster the cement pastes will leach. (2) Hydration kinetics of the accelerated cement are not affected by the w/c. The AS-accelerated cement pastes with lower w/c have a denser pore structure. So, the reduction in the w/c contributes to leaching resistance.

Prediction and Rational Design of Stacking Fault Energy of Austenitic Alloys Based on Interpretable Machine Learning and Chemical Composition

“…In order to investigate the fundamental factors influencing SFE, atomic features were incorporated alongside the original compositional features to construct the input features for the machine learning model, [66][67][68] as depicted in Table 2. These additional features encompass various aspects including atom mass, radius, and electron distribution, enabling a more comprehensive analysis of their impact on SFE at the atomic scale.…”

“…Consequently, it becomes necessary to fine-tune these hyperparameters to discover the optimal combination. Bayesian optimization [66][67][68]80] is a widely used technique for optimizing black-box functions. It achieves the optimal solution by continuously exploring and exploiting the function's information.…”