Türeyen, Erkan BuğraKarpat, YiğitÇakmakcı, Melih2018-04-122018-04-1220160743-1619http://hdl.handle.net/11693/37488Date of Conference: 6-8 July 2016Conference Name: 2016 American Control Conference, ACC 2016Additive manufacturing has become increasingly popular for a wide range of applications in recent years. Micro-stereolithography (μSLA) is a popular method for obtaining polymer-based parts. Systems using the μSLA approach usually consist of a vertical positioning system, a light source and a container where the component is built gradually as the polymer is cured at the locations where the ultraviolet light is projected. It has been noted that the motion of the positioning system and the intensity of the light source is an important factor to achieve high level dimensional precision. In this paper a three dimensional error based learning scheme is presented to improve the time varying process parameters of the system so that the dimensional accuracy of the product is improved. A mathematical model of the curing process is used for developing the error based learning algorithm. The current process parameters as a function of time and the dimensional error obtained at each layer of the production are used for increasing the quality and precision of the same part in the next iteration. Our initial simulation results show significant improvements can be obtained in a few iterations if the correct learning parameters are used based on the target parts dimensional properties.EnglishMathematical modelResinsCuringShapePolymersLiquidsConference Paper10.1109/ACC.2016.7525020