Show simple item record

dc.contributor.authorCollins, James S.
dc.date.accessioned2018-08-20T15:38:29Z
dc.date.available2018-08-20T15:38:29Z
dc.date.created2018-08
dc.date.issued2018-07-24
dc.date.submittedAugust 2018
dc.identifier.urihttp://hdl.handle.net/1853/60295
dc.description.abstractThe goal of representing the physical state of a part throughout the manufacturing process as a digital twin has become a popular topic in recent years. With new advancements in additive-subtractive (hybrid) technologies, the need to gather and process spatial information from inside a CNC system has intensified. Open-loop g-code execution continues to operate with no feedback to describe the current state of the workpiece. It is evident that scanning sensors must be integrated into future machining systems in order to construct a closed-loop architecture whereby the controller can process geometric data to update subsequent commands. This dynamic, closed-loop g-code architecture will revolutionize manufacturing. In order to advance the research in close-loop machining systems, this thesis presents a simple but novel technique for voxel volume model registration. This is done through the application of registering and machining near-net-shape structures and rough castings. Through the implementation of a Euclidean distance transform and variance calculation, an intensity-based similarity metric is demonstrated over a discrete voxel domain driven by a metaheuristic registration algorithm. Simulation tests conducted over a uniform grid structure show that the technique is successful in positioning a floating volume inside its corresponding near-net-shape. Results for six 0.1mm resolution voxel models are reported followed by the metric’s performance under different starting conditions and registration constraints. Tests indicate that the technique works best for narrow to moderately offset volumes. The technique is presented as a prototype to demonstrate the viability of the method. Further applications and refinements of this simple technique will provide engineers with an additional method for part registration to be used in future developments of closed-loop machining systems.
dc.format.mimetypeapplication/pdf
dc.language.isoen_US
dc.publisherGeorgia Institute of Technology
dc.subjectDigital twin
dc.subjectVolume registration
dc.subjectImage registration
dc.subjectCAM
dc.subjectVoxel
dc.subjectClosed-loop CNC
dc.subjectCNC closed-loop machining
dc.titleDigital twin volume registration for Voxel-based closed-loop machining systems
dc.typeThesis
dc.description.degreeM.S.
dc.contributor.departmentMechanical Engineering
thesis.degree.levelMasters
dc.contributor.committeeMemberTucker, Thomas M.
dc.contributor.committeeMemberWang, Yan
dc.date.updated2018-08-20T15:38:29Z


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record