4 Million NOK for fully automated QA loop for 4.0 engineer-to-order manufacturing
CMR scientist Ove Daae Lampe and collegues from IFE and Møreforsking had great fun for three days. They also succeded in landing financial support, 4 Mill NOK "sandpit" funding for this project.
The dreamteam: from left Ove Daae Lampe (Christian Michelsen Research AS); Mikhail Shlopak (Møreforsking Molde AS); Øyvind Jensen (Institutt for energiteknikk, IFE) (Photo: Ove Lampe)
The Norwegian Research Council recently invited to IDEALAB "Industry 4.0 in norwegian". This is the third time Idélab is arranged and 28 researchers and users sat together for three days focusing on digitalisation of the industry. Ove and his colleagues idea is this:
We will make machine learning for quality assurance (QA) available for small series of customized products. In large-scale series production, machine learning for QA is routinely used. For Norwegian SMEs specializing in small series and engineer-to-order production, such an automatic QA is not available. The series are too small to be able to train the machine learning system before the company can move on to the next product.
We will create a machine learning system built around a component that can simulate the unique properties of the part being manufactured. This system is fed with real-time data from the manufacturing process so that the simulation represents what is physically occurring during the production. This way, specific information about the manufactured part is included in the machine learning system. When the company creates the next series, a new relevant simulation component will be used, but the machine learning system will persist. The machine learning system will then collect knowledge that is general and independent of the particularities of the manufactured parts.