Lead Institution:  University of Huddersfield
Partner Institution:  University of Nottingham
Industrial Partners:  Zeeko LtdGlyndwr Innovations Ltd
Project Team:  Prof David Walker (PI, Huddersfield), Prof Sanja Petrovic (Nottingham), Dr Andrew Longstaff (Huddersfield), Dr Simon Parkinson (Huddersfield), Prof Paul Ward (Huddersfield), Dr Wenchang Pan (Huddersfield) and Dr Kyle Wilson (Huddersfield)
Project Duration:  12 months (01 April 2017 – 31 March 2018)

Information Sheet

Research Challenge
The manufacture of precision and ultra-precision functional surfaces for optical, medical and engineering products embraces many artefacts (e.g. optics, prosthetic joint implants, moulds and dies) and several strategically-important sectors, from consumer products, through healthcare, aerospace and automotive, to defence and space.  It is remarkable just how much craft-based hand-work is still used in these extremely high value industries, e.g. had-lapping of optics, precision moulds & dies etc. Whilst this is progressively giving way to CNC machines, these are still highly-dependent on craft expertise to optimise the results. This may involve know-how in special techniques for different materials and surface-forms. Then, crafts-input will be required in interpreting measurement results, and then for planning the right approach to remedy particular surface features that arise, or resolve unexpected process anomalies. Unfortunately,
this know-how is being permanently lost as expert craftspeople retire.

This project has a vision of an adaptive Autonomous Manufacturing Cell for iterative precision surface-fabrication. Artificial Intelligence (AI) is the next key step in practically reducing and eventually eliminating human interventions, both in planning the initial strategy for processing some specific industrial component, and then in interpreting measurement data and adaptively planning each successive process-step. AI should also recognise and respond to unexpected process events, and finally determine when the part has reached specification. To achieve this, this will require a basic feedstock of digitally-accessible, high-quality information. This feasibility study is an essential building block for the ultimate vision of the autonomous manufacturing cell.

The aim of this project is to establish the feasibility, and a workable methodology, for capturing and encoding craft-expertise in a real-life iterative manufacturing context, then to define how optimally to archive it and present it as an accessible input to a future AI system within a future autonomous manufacturing cell.