Senior Capstone
For my senior capstone project, I worked on a powder control system for cold spray 3D printing. I worked in a team of 5 people to develop a flow diagram, a flow diversion system, and a powder containment unit. I was responsible for the design of a custom valve actuator setup and the electronic controls for the system.
The Problem:
During cold spray 3D printing, there is no way to quickly turn on/off the powder feeder during a print which causes over deposition in the parts.
Additionally, the supersonic nozzle requires both a heated gas line and a gas + powder line at specific volumetric flow rates at all times to keep steady flow.
Solution
An actuation mechanism was created in SolidWorks to actuate 2 valves in <.1 seconds, one with the gas line and one with the gas + powder line. When the gas + powder line is diverted away, a purely gas line is substituted in at the same time to keep the volumetric flow at the nozzle constant. The diverted line is sent to a powder collection unit. This allows the powder flow to start and stop rapidly at the nozzle.
The gearbox was designed to be modular to allow for different power sources/ratios to be used during testing. Several parts were redesigned to be waterjet/3D printed to minimize in house manufacturing times.
Note: Some hardware/spacers/supports are not shown for model clarity.
Mechanisms
The impact wrench was controlled remotely via a set of relays and an Arduino. The Arduino also took in sensor data from hall sensors and an absolute rotary encoder attached to the central shaft. Using these, the valve could precisely be moved between positions (+90° and -90°). The spring stop was originally a hard stop, but if the controls stopped working, it would shear the gear teeth if the wrench kept torqueing past the set endpoints. The spring stop engages at the ends to stop excess motion and allows full rotation in the event of control failure.
The Arduino also takes in a 24V signal from the robotic arm to switch states of the powder flow.
Analysis
Stress and fatigue calculations were performed for all major components to ensure safety during operation. This was performed either by hand or in SolidWorks FEA.
Results/Future Improvement
The setup was tested and was successfully able to divert powder away and to the nozzle at will. Due to long lead times with a mass flow controller, a full test with printing was not able to be performed. This prototype proved the validity of this control scheme and provides a base to build a final version off of.
The actuation time was 0.4s for 180° which is higher than the 0.1s goal. This was still faster than off the shelf components that were tested and it is believed that the target time can be achieved through gear ratio optimization. Due to time constraints, only one ratio and power source was able to be tested.
Particle flow was verified through a velocimeter and the results can be seen below.
