Optimizing 3D-printed cement flow through multiscale modeling

Conventional construction in disaster-struck areas, military applications, or other extreme environments is often difficult or unsafe for workers to achieve. Using 3D cement printing offsite and bringing the finished pieces to the location is an excellent alternative. However, there is a gap in understanding what controls the flow behavior of wet, freshly mixed cement, limiting the ability to optimize building performance.

Pal et al. use multiple models to predict how wet cement properties change under different parameters. Their work identified a reduced viscosity with an increased water-to-cement ratio. By comparing their simulations with experiments using wet cement, they observed the same key qualitative trends.

“We first calculated how water affected the interaction between cement surfaces at the atomic scale and then used that information to construct an effective interaction for a much larger particle-based model,” said author Subhadeep Pal.

He said that their multiscale modeling approach is like using a microscope to look at cement at different levels.

By looking beyond the viscosity curve and focusing on what cement particles are doing during flow, the authors could propose a microscopic mechanism for the observed shear thinning.

“I hope this work provides a useful framework for connecting molecular-scale cement science with the particle-scale behavior that ultimately controls processing,” Pal said.

Cement is a complex material with irregular particles. The team’s simplified model was a starting point to identify isolated relationships between water content, particle interactions, shear, and viscosity. Future work will add more elements to the models to ultimately better understand cement flow and use in 3D printing.

Source: “Molecular investigation of freshly made cement-paste rheology for 3D-printing application” by Subhadeep Pal, Shady Gomaa, Gianluca Cusatis, and Sinan Keten, Journal of Applied Physics (2026). The article can be accessed at https://doi.org/10.1063/5.0332157 .