Estimation of the ionic diffusivity of virtual cement paste by random walk algorithm

In order to study the ionic diffusivity of cement-based materials, a numerical model is proposed based on cement hydration and a microstructure model HYMOSTRUC3D, coupled with random walk algorithm (RWA). By computer modeling of random movements of virtual ions in the pore structure of cement paste from the HYMOSTRUC3D model, the ionic diffusivity of cement paste can be obtained according to statistical data analysis. Einstein–Smoluchowski equation which provides the intrinsic relation between the macroscopic diffusion process and the ions Brownian motion under microscopic scale is utilized in the prediction of the ionic diffusivity of cement paste. The proposed model is first validated by experimental results of chloride diffusion coefficient from literature. Simulations show good agreement with the experiments from steady-state diffusion cell test. Next, the proposed model is used to predict the ionic diffusivities of cement pastes with different water-to-cement ratios (w/c) at various hydration times. Then, considering the pore structure of cement paste is an essential part for analyzing the ionic diffusivity of cement paste, the reliability of the pore structures from the HYMOSTRUC3D model is discussed by comparing the pore size distributions by simulation with those by experimental techniques. Finally, the simulated ionic diffusivities of cement pastes in this study are compared with other simulations from literature.

► We propose a numerical model to estimate the ionic diffusivity of cement paste.
► We demonstrates how to estimate the ionic diffusivity from micro-scale to macro-scale.
► The model is validated by experimentally obtained chloride diffusion coefficient.
► Diffusivity increases with increasing w/c and decreases with increasing hydration degree.
► The pore structure by simulation shows good reliability by comparing to experiments.