Owing to the emergence of multi-axis printers, such as robotic printers, curved layer fused deposition modelling becomes physically realizable, which is able to eliminate most of the stair-step effect by continuously adjusting the nozzle orientation and allowing non-planar freeform printing layers that can realize support-free printing (at least in theory). However, robotic curved layer additive manufacturing faces two major issues: 1) When a printing layer is concave (i.e., its curvature vector points outward at the printing point), the potential nozzle-layer local gouging becomes a serious concern, which is especially problematic for metal printing wherein the size of the nozzle head is usually large; 2) the robotic manipulator is an open serial structure, the shape and pattern of a p...[ Read more ]

Owing to the emergence of multi-axis printers, such as robotic printers, curved layer fused deposition modelling becomes physically realizable, which is able to eliminate most of the stair-step effect by continuously adjusting the nozzle orientation and allowing non-planar freeform printing layers that can realize support-free printing (at least in theory). However, robotic curved layer additive manufacturing faces two major issues: 1) When a printing layer is concave (i.e., its curvature vector points outward at the printing point), the potential nozzle-layer local gouging becomes a serious concern, which is especially problematic for metal printing wherein the size of the nozzle head is usually large; 2) the robotic manipulator is an open serial structure, the shape and pattern of a printing path and the post processing method now have a huge impact on the mechanical behavior (e.g., kinematic performance) of the printer, which in turn would lead to poor printing quality and prolonged printing time. In this thesis, we propose a complete pipeline for robotic printing aiming at solving the two major problems.

First, a novel curved layer decomposition algorithm based on the use of (only the positive sides of) ellipsoidal surfaces is proposed to avoid the potential nozzle-layer gouging problem. To facilitate the generation of ellipsoidal layers for an arbitrary model (even for those have a non-zero genus number). The proposed ellipsoidal decomposition algorithm makes use of the skeleton to first decompose a given model into a set of segments (i.e., sub-parts), each of which is bounded by two spheroidal patches. Furthermore, we will generate the ellipsoidal printing layers for each segment by adaptively adjusted the parameters of ellipsoid to satisfy the necessary constraints such as the cusp-height threshold and the support-free requirement. Next, this thesis considers the task of how to generate a printing trajectory that is both free of singularity and having the desired high time-efficiency and good kinematic performance. Specifically, a potential field based algorithm for automatically generating a multi-axis printing path for an arbitrary free-form curved layer in a robotic curved layer printing process is presented. The potential field is useful to identify the local optimal feed directions on the curved layer aiming at improving the printing efficiency and joints' kinematics, and then the printing path is generated to best fit the optimal feed directions by employing the well-known iso-cusp height expansion method.

Finally, to smooth the sharp-changing undulations (induced by the inevitable numerical errors) of the end effector trajectory, this thesis presents a novel B-spline smoothing method for the end-effector trajectory of robot arm subject to the required cusp-height threshold on the printed surface lest the printer would chatter or vibrate and seriously jeopardize the printing quality. In addition, we propose a feed rate scheduling strategy that will try to maximize the variable feed rate while subject to the kinematic constraints of the six joints of the robot arm.

Preliminary tests in both computer simulation and physical printing experiments of the proposed algorithms are conducted, and the results give a positive validation on the effectiveness and feasibility of the proposed methodology.