An intelligent move

22 March, 2017

IPT makes the point that, currently, the norm is for machines to produce parts in networked, pre-programmed production runs: pieces are turned, milled and measured in a set order. But what happens when a machine fails or a customer changes its order? Production has to be re-configured, which is time-consuming and expensive. With this in mind, IPT is looking at an alternative methodology. Instead of a central control program issuing commands, the workflow would develop flexibly, each part deciding for itself the best route through production.

This may sound fanciful and an idea that is far from being realised. However, developers in Fraunhofer are developing just such a system. It is called ‘Service-Oriented Architecture for Adaptive and Networked Production’ and functions similar to an automobile navigation system that uses current data to determine the best route in real time. Each part carries information regarding the next production stage; which machine will be called into operation is purposely left undecided. Only when a production stage is pending does the system select a machine from those that are readily available. Each part bears a QR-Code identifying it as a unique entity.

IPT explains that the software remembers what was done to each part at each production stage; for example, ‘Hole is drilled with machine parameter A and tool X’. A digital twin emerges from this history, displaying at any time where its physical counterpart is in the production process. The Institute explains that digital twins are especially valuable to manufacturers of a wide variety of goods because updating or changing a production run does not require a system overhaul. The ‘Smart Manufacturing Network’ manages the digital twin, always analysing and reusing its process data to improve process robustness and product quality. Michael Kulik, project leader at Fraunhofer working on the software development, said that networking machines with parts will enable companies to produce one-off products in the future – production runs of one.

Kulik explained that a unique aspect of the system is the menu which configures a production sequence. Using drag-and-drop, the user selects individual steps from a list of all services and arranges them in the desired order like building blocks. If a machine fails, a part is simply rerouted to another available machine. “Many machines in a production line can perform a variety of tasks,” he said. “For example, a sophisticated 5-axis milling machine can also do the job of a simpler 3-axis milling machine. In the future, the Smart Manufacturing Network’s service-oriented software can flexibly decide to do the job on an idle 5-axis machine."

Also important for flexible production: Machines from various manufacturers must easily integrate into the Smart Manufacturing Network. IPT is working on this with partners from science and industry in Fraunhofer’s ‘Networked, Adaptive Production’ performance centre. “The plug-and-play that we know from everyday technology does not yet exist in industry,” said Dr. Thomas Bobek, coordinator of the Fraunhofer performance centre. “Our goal is to make plug-and-produce possible.” Fraunhofer’s researchers will demonstrate how the digital twin, service-oriented software and Smart Manufacturing Network collaborate at Hannover Messe as part of the trade fair ‘Research & Technology’ in Hall 2.