Directed energy deposition is an additive manufacturing process in which a line of solid metal is formed by concurrently delivering feedstock material and concentrated energy onto the substrate surface. The process analyzed in this Thesis involves pneumatically transporting fine metal powder from an initial storage hopper to a coaxial nozzle to form a powder cone around the focal region of a high-powered laser. A region of the substrate and some of the powder is melted by the laser to form a meltpool, which solidifies into the deposition line as the nozzle moves across the substrate. Deposition line height is related to powder flow rate fluence and width is related to laser energy fluence. The line dimensions are usually inconsistent due to accelerations and decelerations of the laser head, heat accumulation, inconsistent powder delivery, and more unanticipated issues. These errors compound over multiple layers, which leads to inefficient material addition rates (MAR) or a full failure of the deposition.

So far in the ARMS (Advanced Research for Manufacturing Systems) Laboratory, a height controller has been developed using a charge-coupled device camera sensor to send height feedback to a proportional integral derivative (PID) controller. The PID controller changes the PFR delivered to the nozzle through a Dynamic Powder Splitter System (DPSS). The goal of the height controller is intralayer height control. However, delays in the PFR response time limit its resolution due to the extended travel distances of the nozzle between erroneous measurements and the arrival of the corrected PFR. In this Thesis, two approaches were taken to reduce the response time as much as possible. First, the DPSS-based powder conveyance system parameters were optimized for PFR response though a full factorial experiment. Then a closed-loop system was instituted by sending PFR feedback via an optoelectronic sensor. PI and PID controllers were developed to correct for PFR errors and purposefully induce overshoot to gain a relationship to PFR response time. Responses to PFR step inputs and single-layer thin wall depositions were used to analyze the open-loop parameter optimization and the closed-loop’s effectiveness at reducing PFR response time. Parameter optimization showed a small improvement to PFR response time, but the results were difficult to replicate during deposition experiments. The closed-loop system with a PI controller was highly effective at reducing PFR response time. The deposition tests with noticeable overshoot revealed a constant relationship between PFR change and travel distance. With sacrificing response time to reduce overshoot, the closed-loop PFR system can be nested into an overarching height controller for faster and more consistent height control. This will be useful for future additive manufacturing applications at high MAR.