Extrusion Process Simulation and Layer Shape Prediction during 3D-Concrete-Printing Using The Particle Finite Element Method

Abstract

To enable purposeful design and implementation of automated concrete technologies, precise assessment and prediction of the complex material flow at various stages of the process chain are necessary. This paper investigates the intermediate stage of the extrusion and deposition processes in extrusion-based 3D-concrete-printing, using a numerical model based on the Particle Finite Element Method (PFEM). In PFEM, due to the Lagrangian description of motion, remeshing algorithms and the alpha shape method are used to track the free surface during large deformation scenarios. The Bingham constitutive model was used for describing the rheological behaviour of fresh concrete. This model is validated by comparing the numerically predicted layer geometries with those obtained from laboratory 3D printing experiments. Extensive parametric studies were then conducted using the numerical simulation, delineating the influence of process and material parameters on the layer geometries, the dynamic surface forces generated under the extrusion nozzle and the inter-layer interactions.

A mixed u–p edge-based smoothed particle finite element formulation for viscous flow simulations

“…This implies that already a coarse spatial discretization is able to predict the expected cross section with sufficient accuracy. The model results obtained for the finest mesh discretization are validated by comparing the printed cross-section with measurements from laboratory tests and with results from numerical simulations using the regularized Bingham model, both taken from [64]. Figure 24 shows the shapes of the printed concrete after 10 s. Almost no difference is found in comparison with the regularized Bingham model, and only marginal differences are observed when comparing the numerical predictions and the experimental data.…”

“…In concrete printing, the fresh, flowable concrete is extruded layer-wise to form structural components without any use of additional scaffolding support during casting [62,63]. It was shown in [63][64][65] that the material deformation of fresh concrete during extrusion can be modeled with sufficient accuracy using viscoplastic consti- 21 Slump flow test: final shape of the slump computed with the finest mesh size at t = 60 s and compared with the analytical solution [33] tutive models. The conducted virtual experiment is based on a one layer test documented in [64].…”

“…It was shown in [63][64][65] that the material deformation of fresh concrete during extrusion can be modeled with sufficient accuracy using viscoplastic consti- 21 Slump flow test: final shape of the slump computed with the finest mesh size at t = 60 s and compared with the analytical solution [33] tutive models. The conducted virtual experiment is based on a one layer test documented in [64]. The material was printed with a round extrusion nozzle with a diameter of 5.6 cm, a printing speed of 0.05 m/s, a distance between the extrusion nozzle and the ground of 2.4 cm and a printed layer length of 50 cm, see Fig.…”

“…Furthermore, a no-slip condition was applied to the interface between the fresh concrete and the ground surface. Based on a previous numerical-experimental analysis [64], typical material properties of 3D-printed fresh concrete were adopted as follows: μ = 5 Pas, τ 0 = 306 Pa, ρ = 2078 kg/m 3 , E = 0.1 MPa and ν = 0.3. The time step size was specified as Δt = 2 • 10 −4 s. The material at the extrusion nozzle was approximated with three different finite element discretizations using a node spacing of approximately 11, 15 and 19 number of nodes along the diameter, which resulted in approximately 16,000, 80,000 and 184,000 finite elements after t = 10 s.…”

“…Fig.243D simulation of concrete printing: simulated cross section after t = 10 s. Comparison between the proposed ES-PFEM model and the regularized Bingham model and experiments[64] …”