Abstract

Implementing a virtual production system is challenging. Production systems are constrained by different dynamical market changes and characterised by complex general and machine specific interactions, uncertainties and unknowns. These complications make predicting the behaviour by any deterministic computer model not only inaccurate but sometimes misleading, e.g. if the range of a model’s applicability is not sufficiently well identified. Thus it is crucial to implement a virtual production system that supports learning from data and existing knowledge in the iterative design of models and their range of applicability – such a systematic design approach integrates real versus virtual data with digital versus human creativity and intelligence. The theory of design oriented thinking adapted for manufacturing favours fast iteration in digitised design cycles instead of optimisation of model quality in one step.

© 2017 Optical Society of America

PDF Article