Project Description

The goal of this research project is to integrate data from various companies and processes and to comprehensively optimize the paper recycling cycle using big data approaches and selected AI tools. A prototype AI application is being further developed to optimize the recycling loop—from the collection and sorting of waste paper to the production of the final paper product—thereby making a significant contribution to reducing resource consumption. The basis for this is, first, improved identification and characterization of the composition of waste paper and, second, building on this, the use of intelligent systems that utilize this information in subsequent stages of the value chain to ensure efficient resource use.

To achieve this goal, a forecasting system is required that, trained on historical data, recognizes based on the current situation when the operator must intervene to make adjustments (decision support and alert function). An assistance system that links forecasts and the current status with expert knowledge generates recommendations for actions, from which the operator selects an option and then takes corrective action (recommendation function). This allows the machine operator to be alerted early on to impending deviations in the characteristics of the finished product or disruptions in the production process, so that the operator can take countermeasures in a timely manner. The assistance system continuously learns from the operator’s selections and their resulting impact on the process parameters (self-optimization function). To implement this, data is exchanged automatically across companies. On behalf of the IPEM Chair, our group leader Lukas Baeck
is serving as project manager.