Concrete is a heterogeneous material with a disordered material morphology that strongly governs the behaviour of the material. In this contribution, we present a computational tool called the Concrete Mesostructure Generator (CMG) for the generation of ultra-realistic virtual concrete morphologies for mesoscale and multiscale computational modelling and the simulation of concrete. Given an aggregate size distribution, realistic generic concrete aggregates are generated by a sequential reduction of a cuboid to generate a polyhedron with multiple faces. Thereafter, concave depressions are introduced in the polyhedron using Gaussian surfaces. The generated aggregates are assembled into the mesostructure using a hierarchic random sequential adsorption algorithm. The virtual mesostructures are first calibrated using laboratory measurements of aggregate distributions. The model is validated by comparing the elastic properties obtained from laboratory testing of concrete specimens with the elastic properties obtained using computational homogenisation of virtual concrete mesostructures. Finally, a 3D-convolutional neural network is trained to directly generate elastic properties from voxel data.

Computational models that predict melt pool shapes and temperature evolution in laser-based powder bed fusion of metals (PBF-LB/M) can range from simple thermal simulation models to more advanced models that incorporate more detailed physics of the process. While advanced models can accurately predict thermal fields and melt pool fluid dynamics, they are computationally more expensive and, thus, less suited for part-scale simulations or numerical optimization, where repeated model evaluations are necessary. On the other hand, thermal simulations are computationally efficient and attractive for their simplicity, but their accuracy is mainly limited to conduction-dominated processes. Moreover, the conduction model’s validity range is not fully understood for non-Gaussian laser beam shapes. This paper demonstrates that predictions of melt pool depth and width carried out by a heat conduction model are accurate to within 20 % for all investigated laser profiles, provided that the simulated maximum temperature does not exceed a certain threshold value for stainless steel 316L. This is established by thoroughly investigating the validity range of the heat conduction model through comparisons with over 200 single-track experiments on bare plates. The temperature predictions from the model are compared with multi-physics simulations using smoothed particle hydrodynamics (SPH). Through detailed analysis and validation for three laser beam shapes, this contribution provides valuable insights into the accuracy and applicability of heat conduction models in bead-on-plate melting simulations and offers a path to optimize process parameters, such as laser beam shape, scanning strategy, and other processes for diverse applications aimed at PBF-LB/M.

Concrete is a heterogeneous material with a disordered material morphology that strongly governs the behavior of the material. In this contribution, we present a computational tool called the Concrete Mesostructure Generator (CMG) for the generation of ultra-realistic virtual concrete morphologies for mesoscale and multiscale computational modeling and simulation of concrete. Given an aggregate size distribution, realistic generic concrete aggregates are generated by a sequential reduction of a cuboid to generate a polyhedron with multiple faces. Thereafter, concave depressions are introduced in the polyhedron using gaussian surfaces. The generated aggregates are assembled into the mesostructure using a hierarchic random sequential adsorption algorithm. The virtual mesostructures are first calibrated using laboratory measurements of aggregate distributions. The model is validated by comparing the elastic properties obtained from laboratory testing of concrete specimens with the elastic properties obtained using computational homogenisation of virtual concrete mesostructures. Finally, a 3D-convolutional neural network is trained to directly generate elastic properties from voxel data.

The paper at hand evaluates the validity of the transient heat equation with phase change and temperature-dependent coefficients as a model to predict the evolution of melt pools for rapid turnaround scan strategies in a laser powder bed fusion (PBF-LB) additive manufacturing (AM) process. In such scenarios, residual heat effects due to short laser-off times between tracks significantly impact the melt pool morphology. Results show that the conduction model accurately captures the rapidly evolving melt pool morphologies under these conditions. This is a fundamental discovery, as the ability to predict such dynamic scenarios has previously only been attributed to much more involved models which include fluid-like effects that otherwise require computational fluid dynamics (CFD). However, the presented validation (e.g., validated comparison to surface topography measurements) clearly demonstrates that a large portion of the observed effects is actually captured by the heat diffusion equation, including a common model of phase change. As a proof of concept for an application of this insight, feedforward control is investigated. Therein, the validated model predicts the melt pool morphology for a modified scan strategy with an increased pause time between the tracks that reduces melt pool distortions observed in the original configuration.