Generative MOR-Based Design of Modular Carbon-Reinforced Concrete Structures

  • Macek, Domen (RWTH Aachen University)
  • Döpke, Finn (RWTH Aachen University)
  • Brepols, Tim (RWTH Aachen University)
  • Holthusen, Hagen (Friedrich-Alexander-Universität Erlangen-Nürn)

Please login to view abstract download link

In this work, we present a generative design and analysis framework for modular, slender carbon-reinforced concrete (CRC) structures with variable component geometries (e.g., [1]). The primary objective is to enhance structural performance while enabling a more efficient and sustainable use of materials. The proposed framework integrates an automated structure generator with a highly efficient parametric analyser based on model order reduction (MOR) technique [2]. The structure generator systematically creates complete modular systems by varying geometric parameters, loading conditions, material assignments, and spatial material distributions. Different generation strategies are proposed, including optimization-based approaches for the systematic exploration of geometric and modular design parameters and data-informed algorithmic approaches for identifying promising structural configurations. Due to the geometric complexity and large number of generated designs, direct finite element analysis would be computationally expensive. Therefore, the MOR-based parametric analyser exploits the modular architecture by reducing substructures independently and efficiently assembling them into global reduced-order models. This enables rapid mechanical evaluation and verification of structural constraints across extensive design spaces. An iterative information exchange between the structure generator and the parametric analyser allows evaluation results to continuously guide subsequent design generation. The framework automatically compares and ranks candidate structures according to predefined mechanical and sustainability-related performance criteria. The resulting design tool produces multiple optimized CRC structural configurations tailored to specific boundary conditions and functional requirements, demonstrating the potential of combining generative design, and reduced-order modelling for scalable and resource-efficient structural engineering. REFERENCES [1] Goertzen T., Neef T., Scheffler P., Macek D., Mechtcherine V., Niemeyer A. C., 3D concrete printing of topological interlocking blocks, Materials & Design, Vol. 254, Art. 114049, 2025. [2] Ritzert, S., Macek, D., Simon, J.-W., Reese, S. An adaptive model order reduction tech- nique for parameter-dependent modular structures, Computational Mechanics. 2023, 1–17. https://doi.org/10.1007/s00466-023-02404-w