Mechatronic Design and Feedback Control for Quality-assured Laser Powder Bed Fusion Additive Manufacturing

Thesis (Ph.D.)--University of Washington, 2021Despite continuous technological enhancements and emerging applications of laser powder bed fusion (LPBF) additive manufacturing, the lack of repeatability and stability still represents a barrier to industrial breakthrough. This dissertation developed new mechatronic design of process monitoring and capabilities of robust control to reduce process variation and ensure part quality. By building a sensor-rich feedback-centric LPBF platform, this study enables access to crucial process signatures (e.g., powder bed temperature, melting pool image) and manipulation of various process parameters (e.g., heater power, scan velocity) to provide the opportunity for model-based controls. Leveraging the established mechatronic design, we investigated control algorithms and data interpretation frameworks for attenuating mechanical vibrations, thermal transfer fluctuations, and laser-material interaction imperfections. Specifically:We introduced a model-inverse free local loop shaping control for disturbance rejection in the laser scanning process. We proposed and verified filter designs for attenuating both narrow- and broad-band vibrations in the position output of the scanner, leading to higher accuracy of part geometry. We proposed a model-based multi-zone heating control algorithm for minimizing the powder bed temperature deviation in the presence of part-geometry induced cross-layer thermal disturbance, leading to a more homogeneous part mechanical property. We created an image analytics framework for defect identification, part quality evaluation, and process signature filtering, which paves the road towards online implementable laser-material interaction control