The Use of Iterative Learning Control to Improve the Accuracy of Additive Manufacturing

Fused deposition modeling (FDM) additive manufacturing, while versatile, has performance limitations in certain geometries, such as arcs, and holes. Iterative Learning Control (ILC) is good at accounting for disturbances to a system that are constant, but difficult to model and correct. Since the printer is doing repetitive actions, but not producing accurate parts, it is thought that ILC could be used to compensate for the disturbances thus improving accuracy. In order to start the development of the control system it was first necessary to get a baseline for the printer and scanner. Based on the initial results of the testing it was concluded that both the printer and scanner were a source of error and thus would both need to be accounted for by the ILC. This error can be attributed to, on the printer side of things, uneven cooling, manufacturing imperfections, and even the way a G-Code file is made, and, on the scanner side of things, the capture resolution, and alignment imperfections. The ultimate goal of the research and development of the ILC is to allow for inexpensive “RepRap” machines to produce parts with a level of accuracy that is closer to high end industrial FDM machines.