This Faculty Early Career Development Program (CAREER) project will enable substantially higher accuracy and greater reproducibility in additive manufacturing (AM) processes. In contrast to conventional machining, where parts are made by cutting away unwanted material, additive manufacturing -- also called 3D printing -- builds three-dimensional objects of unprecedented complexity by progressively adding small amounts of material. Powder bed fusion (PBF), in which new material is added to the part being fabricated by applying and selectively melting a powdered feedstock, is a popular form of AM for fabricating complex metallic or high-performance polymeric parts. This project supports fundamental research to create new thermal modeling, sensing, and control algorithms that will lead to precise and reliable PBF. The modeling task will enable fast and accurate prediction of heat flow and temperature distribution during powder fusion. The resulting knowledge on directing heat flow is essential for achieving a desired three-dimensional shape. The sensing task will formulate new signal processing algorithms that discard unnecessary information to make full use of data-intensive sensor sources like high-speed video. Finally, these results will be integrated with new control algorithms in order to counteract process variations and provide repeatable, low-cost, high-quality parts. AM offers untapped potential in a wide range of products for the energy, aerospace, automotive, healthcare, and biomedical industries. PBF parts are increasingly preferred in applications ranging from advanced jet-engine components to custom-designed medical implants. Therefore, the outcomes of this project will facilitate fabrication of products to benefit the US economy and improve quality of life. Broader impacts of the project will be augmented by dissemination of educational results via a network of twenty-four collaborating universities, to inculcate skills for innovative problem solving into undergraduate engineering education.
The powder bed fusion process exploits precision heating and rapid solidification, together with layer-by-layer adjustments to feedstock application, and scan speed and path of lasers or electron beams. This project will expand knowledge at the interface of modeling and process controls, to consider the main obstacles to precision manufacturing with AM. Specifically, the project will address (1) the lack of tractable online models that capture multi-scale thermomechanical interactions, and (2) the need for control strategies in the presence of limited-bandwidth sensor feedback. A dynamic real-time model will be produced through separation of the cross-scan and cross-layer dynamics, allowing currently intractable powder fusion dynamics to be treated in real time, using computation-friendly primitives. Then the structure of the process dynamics will be used to enable a feedback controller for laser energy deposition. Controlling the proper energy deposition is critical for ensuring quality and reproducibility. The approach will be based on modeling and adaptation methods originally developed in precision mechatronics, in conjunction with a formulation of multi-rate control that can reject structured thermal disturbances at a fast, user-configurable sampling rate. Collectively, the project will add the needed new knowledge on quality assurance to future repetitive and layer-by-layer thermomechanical processes, by (1) establishing a physics-based, control-oriented modeling approach to understand and engineer the layered thermal interactions, and by (2) creating a foundation for closed-loop control solutions to produce desired uniform temperature fields in periodic and near-periodic deposition of thermal energy.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
