Development of advanced part recognition toolkit for automated additive manufacturing post-processing system

Additive Manufacturing Technologies (AMT) is a Sheffield based manufacturer of smart post-processing systems for additive manufacturing (AM). During recent years, AMT has attracted multi-million-pound investment for the development of automatic surface smoothing, de-powdering and colouring machines. AMT's vision is to create a Digital Manufacturing System (DMS), involving smart metrology and AI controlled post-processing modules.

This project aims to develop geometry measurement and machine learning integration into AMT's DMS system. AMT require an analysis toolkit for parts at various states, including powdered, de-powdered, un-processed, post-processed, as well as various geometries and printing methods. This requires research of appropriate machine learning algorithms able to recognise and categorise the surfaces in real time while using limited computational power. Among the tasks for such machine learning is quality control of the geometry of AM part surfaces, involving internal and difficult-to-access areas.

The machine learning will steer the parameter controls for each module of the DMS. Surface processing databases will be generated for smart automated control of post-processing stages. The process will allow smoothing, colouring, functional coating or surface morphology manipulation by iterative self-learning, making it a game-changing technology for AM surface engineering and digital manufacturing. The process will enable a new generation of AM parts to be used in existing applications, e.g. battery/capacitor assemblies, satellite coatings, antimicrobial medical components and creation of new structural fibres.