A Multi-Input Single-Output iterative learning control for improved material placement in extrusion-based additive manufacturing

Abstract

A major limitation in extrusion-based additive manufacturing (AM) is the lack of process monitoring and control tools in the material deposition frame. Iterative learning control (ILC) has been demonstrated to be an effective control strategy (Hoelzle et al., 2009; Bristow & Alleyne, 2003) due to the repetitive nature of manufacturing processes, but for much of the prior work of ILC in AM, the focus was on precise control of the machine components. To improve on fabricated part quality, we apply ILC directly to the material deposition to account for uncertainty in material behavior and imperfect coordination between the material flow and the machine axes. This paper presents a novel Multi-Input Single-Output (MISO) Iterative Learning Control (ILC) method that couples the extrusion input and axis speed to improve material width control along the trajectory. MISO ILC partitions the error to the different stages through frequency separation to avoid input saturation. MISO ILC uses process feedback in the material deposition frame to improve material placement. The stability and convergence properties for the MISO ILC system are presented in the lifted domain. Simulation and experimental system results on an extrusion printing system demonstrate that MISO ILC achieves improved material placement control along the path. While we apply the MISO ILC to a specific extrusion system, we present a general control strategy that can be implemented for other MISO applications.