Towards a one-shot printed gripper for part extraction from powder-based 3D printing processes

Cormack, Jordan

Abstract

As 3D printing moves from a tool used predominantly for creating one-off prototype parts to being a key part in high volume manufacturing, the need for increased automation becomes desirable to save time and cost in the process. For 3D printing processes which are already used for high volume manufacture, such as HP's Multi Jet Fusion, there is currently no automated method for removing printed objects from the powder based build volume individually, keeping them in a known pose. This thesis details the development of a compliant one-shot 3D printed gripper mechanism which is able to extract individual caked 3D printed objects from aged PA12 powder, in a known pose. The gripping mechanism was simulated using finite element analysis, and validated using a visual tracking tool. Through the use of strain gauges applied to the 3D printed gripper, it was found that the unfused powder does not easily flow from between the tips and the object, causing its perceived size to increase, and instability to occur during the grasp. A novel modified tip geometry and local vibration was shown to completely remove any increase in the perceived object size and slip during extraction from virgin PA12 powder using cleaned objects. Multiple caked objects were also extracted from within aged powder in a known pose, closely representing the actual MJF process, including a high variability of caked powder surrounding the objects. A new sensor comprised of an array of strain gauges was developed based on the toothed tip geometry of the gripper, which is able to provide more rich information about object grasp, such as force, torque, and slip detection. Elements of this work are also applicable more generally for grasping dirty objects, or ones buried in loose media. Future work to replace the commercial sensors and motors with fully 3D printed alternatives will allow the gripper to be manufactured even quicker, for less cost, and allow it to be fully tailored to the 3D printed objects which need extraction.

• Thesis Type: Thesis
• Deposit Date: Mar 15, 2022
• Publicly Available Date: Aug 4, 2022
• Public URL: https://uwe-repository.worktribe.com/output/9199390
• Award Date: Aug 4, 2022