Industry 4.0: Akzent4BaSys funding project

Tuesday, 9:45 am, production hall of GS Kunststofftechnik in Idar-Oberstein. The machines are running at full speed, it is exciting to see how a gripper arm lifts a complex workpiece out of the injection moulding machine every few seconds. We are in the factory of one of the leading medium-sized companies in Germany that process plastics. There is a good reason for this: in addition to GS Kunststofftechnik and three other consortium partners from the fields of research and electronics, we are involved in a pioneering project initiated by the German Federal Ministry of Education and Research (BMBF): Accent4BaSys. The aim is to make GS Kunststofftechnik's production facilities industry 4.0-capable, optimise production processes and maintain machines with foresight. How? With the help of our intelligent NLP speech analysis and the connection of all machines via a standardised BaSys 4.0 middleware. This will soon make it possible for employees to speak the same language among each other, as well as between man and machine, to communicate and exchange knowledge.

Whether the Battenfeld B6 injection molding machine, the handling systems or pressure units - all components in the plastics factory supply different process data via different interfaces and data formats. For each machine, employees can only view a small selection of data on one display - everyone gets the same display, which means: in their different roles and powers, they cannot view additional machine data. GS Kunststofftechnik had introduced a Manufacturing Execution System (MES) some time ago and now wants to connect its machinery to the BaSys 4.0 middleware. However, the middleware lacks a way to flexibly integrate the plant employees on the shop floor and enable them to exchange the converging information regardless of location.

On different levels, the research project at GS Kunststofftechnik brings together what is highly relevant in practice for other medium-sized companies:

• Vesatec: The company's task is to obtain data from the production plants, some of which have been running for over 40 years.
• Kybeidos: We integrate the employees into the processes on the basis of our own VDPP platform via analyzed language.
• German Research Center for Artificial Intelligence (DFKI) and Fraunhofer IESE: The research institutions provide AI know-how as well as the standardized BaSys 4.0 middleware, which integrates speech analysis and sensor data into the existing production system.

At the end, a demonstrator will be created as well as an extension of the BaSys standard that can be universally used in production halls. It ensures that machines do not stop in the best case scenario. Faults are detected at an early stage and eliminated before damage occurs. Should damage occur, the repair process is accelerated. The result: reduced damage and downtimes, fewer stress situations for employees, shorter walking distances and better team transfers. It is expected that the machine utilization and the effectiveness of the entire plant and thus also the cost efficiency will be increased by up to 10 percent.

We analyse the communication content and generate a decisive added value

First we evaluate open source messenger platforms. It is important that the platform used is open enough for us to technically connect our speech analytics and thus our VDPP platform. Of course, it must also meet the data governance and security requirements of GS Kunststofftechnik. The interaction platform will ultimately provide video telephony, embed machine data in a visualized form (e.g. a diagram showing the course of oil pressure) and automatically send messages to ticket systems and the MES as well as to maintenance staff, shift supervisors and production managers.

The highlight: Our technology is industry-neutral, comprehensively evaluates text and speech and can be trained on the technical vocabulary of GS Kunststofftechnik - in the sense of comprehensive knowledge management. What's more, the employee himself controls what information he needs. If desired, he or she can "switch on" the language analysis into the communication. At the right moment, it then gives a hint that ensures that the machine will be running again in two hours rather than two days.