Understanding the transient properties of cementitious pastes is crucial for construction materials engineering. Computational modeling, particularly through Computational Fluid Dynamics (CFD), offers a promising avenue to enhance our understanding of these properties. However, there are several numerical uncertainties that affect the accuracy of the simulations using CFD. This study focuses on evaluating the accuracy of CFD simulations in replicating slump flow tests for cementitious pastes by determining the impact of the numerical setup on the simulation accuracy and evaluates the transient, viscosity-dependent flows for different viscous pastes. Rheological input parameters were sourced from rheometric tests and Herschel-Bulkley regression of flow curves. We assessed spatial and temporal convergence and compared two regularization methods for the rheological model. Our findings reveal that temporal and spatial refinements significantly affected the final test results. Adjustments in simulation setups effectively reduced computational errors to less than four percent compared to experimental outcomes. The Papanastasiou regularization was found to be more accurate than the bi-viscosity model. Employing a slice geometry, rather than a full three-dimensional cone mesh, led to accurate results with decreased computational costs. The analysis of transient flow properties revealed the effect of the paste viscosity on the time- and shear-dependent flow progress. The study provides an enhanced understanding of transient flow patterns in cementitious pastes and presents a refined CFD model for simulating slump flow tests. These advancements contribute to improving the accuracy and efficiency of computational analyses in the field of cement and concrete flow, offering a benchmark for prospective analysis of transient flow cases.
Keywords: CFD; OpenFOAM; fresh cement paste flow; non-Newtonian flow; regularization; rheology.