Electronic Theses and Dissertations for Graduate School
Add My Work
Author Last Name
ANALYTICAL PROCESS MODELING AND NONLINEAR CONTROL OF MELT-POOL HEIGHT AND TEMPERATURE IN DIRECTED ENERGY DEPOSITION
Restricted (Penn State Only)
Doctor of Philosophy
Date of Defense:
June 01, 2017
Qian Wang, Dissertation Advisor
Qian Wang, Committee Chair
Panagiotis Michaleris, Committee Member
Bo Cheng, Committee Member
Ted Reutzel, Outside Member
Directed Energy Deposition
Additive manufacturing is a cutting-edge manufacturing technology which can be applied to many areas, for example bio-fabrication, automotive and aerospace. However, this high-potential manufacturing technology hasn’t been widely applied in industry due to several severe drawbacks, such as the poor accuracy of part geometry and inferior metallurgical properties, which can be addressed by utilizing feedback control system. This dissertation focuses on developing control-oriented model and model-based control system design for melt-pool height and temperature control. In this dissertation, we first construct a reduced-order physics-based nonlinear model of the melt-pool height and temperature for 1-D case of directed energy deposition process, in which only single straight track is deposited on the substrate by LENS system. One main contribution of our proposed model compared with existing literature lies in a novel parameterization of the entire material transfer rate as a function of the major process parameter in manufacturing process. This novel parameterization is derived from the perspective of physics and can improve the accuracy of our model prediction of the steady-state melt-pool geometry. Then system identification are conducted to calibrate the unknown parameters in our 1-D model using experimental data and FEA prediction collected from deposition of Ti-6AL-4V and Inconel 718, followed by validation of 1-D model with calibrated parameters. A LQR controller and nonlinear MPC controller are designed based on this model to control the melt-pool height and temperature by regulating the laser power and scanning speed applied in the manufacturing process. The dynamics of additive manufacturing process varies inter-bead or inter-layer, which motivates the modification of 1-D model to multi-dimensional one. The 1-D model can be extended to multi-dimensional case by incorporating the thermal history in the existing part, which can be characterized approximately by Rosenthal’s solution. Our extended multi-dimensional model is then validated by experimental data obtained from three multi-dimensional examples: thin-wall structure, parallel patch build and L-shape structure. Next a feedback linearization controller is developed based on this multi-dimensional model to track the melt-pool height reference trajectory using laser power.
Login using your Penn State access account to view the paper.