Even in a highly automated production environment, experienced employees are the key to an efficient production process. If that knowledge is missing disruptions in the production can be solved through symptoms but not the actual cause. In return such disruptions or failures consistently re-occur and produce unplanned downtimes as well as a decrease in quality. Considering the demographic change, shortage of skilled workers as well as employee turnover those challenges increase. Digitalization of knowledge in a knowledge data base is a potential solution. However, employees face high pressure during a downtime or failure and finding the correct entries through keyword searching in a knowledge database is challenging. The consequence is employees stop maintaining and using the existing knowledge database.
Our self-learning assistant system MADDOX uses methods of machine learning and analyzes machine data based on similarities in the past. The search algorithm presents connected entries of colleagues in the knowledge database to a tablet and motivates through ‘confirm’, ‘reject’, correcting and adjusting those entries. Furthermore, the proactive assistance helps to increase the user acceptance and employees are motivated to maintain the system on ahigh level. The innovative communication platform decreases operating costs, increases the efficiency in the production as well as engagement and appreciation and improves documentation.
As part of research and development at the Fraunhofer IVV in Dresden we were able to develop MADDOX in corporation with psychologists in a timeframe of 5 years. These research outcomes helped founding the spin-off Peerox GmbH with a marketable product. By now MADDOX runs at the main factory of the BAYER AG in Leverkusen on several packaging lines.