Besides the information apparent at a first glance, correlations can be extracted from process data, which can support the understanding of the behaviour of processes and plants. These relations can be identified via AI supported information retrieval from process conditions and product characteristics from measured data, i.e. time series and image data in laboratory and production. Thereby, uncertainties about the actual operation regime shall be reduced, interruptions or erroneous trends shall be early identified, and systematic deviations shall be automatically diagnosed and considered for the process control. This enhances the economic efficiency of chemical and biotechnical production processes by the optimal usage of secure operating windows, higher plant availability, and process and plant security.
Contact: Prof. Leon Urbas, TU Dresden
P2O-Lab and ZIH of the TU Dresden