No abstract

No abstract

Additive manufacturing in large-scale construction is an ongoing research topic that shows significant potential to overcome the challenges in terms of efficient material usage and process automation in construction. A large challenge in deposition based additive manufacturing processes of concrete material is to ensure the structural stability while printing. Due to the weak material properties of the fresh concrete, it has to be ensured that during the printing process the not fully cured printed structure is able to carry its own weight. This requires process stabilization and a proper process control to prevent a collapse of the structure. Therefore, a numerical model of the printing process that takes into account the time dependent material behavior of the applied concrete as well as the printing path and the process parameters is necessary to support the process planning. Within the framework of project B04 of the collaborative research center TRR277 -Additive Manufacturing in Construction, a novel path-based finite element simulation was developed in which the simulated geometry is constructed directly from the printing trajectory. Additionally, this approach allows the time-dependent material properties of fresh concrete to be modeled directly and efficiently into the mesh of the printed structure. Since the computation of large scale printing processes with finite element simulations is quite expensive, there exists a need for a much faster computational model. In this contribution, the implementation of a surrogate model based on a neural network and its deployment to optimize the interlayer waiting time is presented.