Injection 3D Concrete Printing: A framework for integrating structural design and automated robotic fabrication of space trusses

Emerging in the latter half of the twentieth century, computational form-finding methods have significantly enhanced structural efficiency and design flexibility. In recent decades, their evolution has increasingly converged with fabrication techniques, highlighting their profound impact on the construction industry. Particularly in the past two decades, the rise of digital fabrication has further revolutionized construction, bringing both new opportunities and challenges in material application and fabrication. However, the integration of design and fabrication still predominantly follows a sequential workflow, which is often tedious and highly iterative. This linear approach frequently necessitates substantial revisions to the initial concept to accommodate constraints at various stages, leading to inefficiencies and project delays. This dissertation is motivated by the potential to integrate structure- and fabrication informed constraints into the early design phase to enhance efficiency across the entire process, from design to fabrication, and to automate the fabrication based on the design input. The research focuses on Injection 3D Concrete Printing (I3DCP), an innovative technique enabling the rapid creation of intricate concrete networks. By incorporating a carrier liquid, I3DCP significantly extends design flexibility compared to conventional horizontal extrusion printing, as it mitigates the effects of gravity. However, challenges arise from material physics, complex geometry and robotic kinematics, making it particularly difficult to automate the generation of feasible motion trajectories for fabrication. This dissertation aims to develop an integrated framework for I3DCP, ensuring controlled workflow management from conceptual design to robotic fabrication while accounting for material behavior, geometrical constraints, and robotic kinematics. Grounded on the assumption that concrete struts in I3DCP behave as pin-jointed elements, the designed structures can be modeled as vector-based space trusses during the conceptual design phase. With predefined material properties, topological and geometrical constraints are derived from preliminary empirical tests. Vector-based Graphic Statics (VGS), a graphical form-finding method, is employed for structural design and optimization. The vector-based model in VGS encapsulates all topological and geometrical constraints, along with internal force values of concrete struts, enabling designers to interactively manipulate and refine structures in real time. For fabrication, the design models are treated as discrete spatial structures to compute the robotic motion trajectories. The assembly sequence is determined by solving a discrete Constraint Satisfaction Problem (CSP), ensuring feasibility under given constraints. The motion trajectory is then generated based on this sequence by optimizing for the geometric and physical limitations of the robotic arm. By consistently incorporating all the aforementioned constraints, the proposed framework guarantees that any structural model input from the design phase has at least one feasible motion trajectory. The framework’s potential is demonstrated through a series of experimental case studies at various scales, including sculptures, a table, and a bridge.