Date: Friday, February 27th, 2026 

Time: 2:00PM - 3:30PM ET 

Location: GTMI 114

Virtual Link: Neel's PhD Proposal - Multimaterial Mass Flow Control [In-person] | Meeting-Join | Microsoft Teams

Meeting ID: 295 197 015 651 40

Passcode: SE6mK3nD

Neel Shah 

Ph.D. Robotics Student

George W. Woodruff School of Mechanical Engineering
Georgia Institute of Technology

Committee: 

Dr. Aaron Stebner (Advisor)

Mechanical Engineering

Georgia Institute of Technology

Dr. Emmanouil Tentzeris

Electrical Engineering

Georgia Institute of Technology

Dr. Levi Wood

Mechanical Engineering

Georgia Institute of Technology

Dr. Anirbar Mazumdar

Aerospace Engineering

Georgia Institute of Technology

Dr. Shreyas Kousik

Mechanical Engineering

Georgia Institute of Technology

Dr. Zachary Brunson

Mechanical Engineering

Georgia Institute of Technology

Dr. Jin Yeon Kim

Mechanical Engineering

Georgia Institute of Technology

Abstract: Functionally Graded Materials (FGMs) offer a solution to complex engineering design trade-offs by enabling the continuous variation of material properties within a single structure. Powder-Blown Directed Energy Deposition (PB-DED) has emerged as a premier additive manufacturing technology for FGM fabrication due to its ability to dynamically alter material composition in-situ. Despite this potential, the process remains difficult to implement in critical production environments due to inconsistent part quality and the stochastic nature of powder flow. Current state-of-the-art systems often rely on basic open-loop control which fail to account for process volatilities such as oscillatory flow instability and hysteresis. This proposal focuses on developing the sensing capabilities and control architectures necessary to bridge the gap toward repeatable manufacture of functionally graded structures. The proposed work is organized into four primary objectives: the development of novel PB-DED mass flow sensing modalities; the implementation of an adaptive feedforward control architecture integrating an MRAC-enhanced Smith Predictor; real-time multi-material flow quantification via spectral sensing; and the implementation of machine learning to distill spectral multi-material flow information into an array of inexpensive, orthogonal-modality sensors.