COGNITWIN: Cognitive Factories via Proactive, Self-learning Hybrid Digital Twins

Short description of the project

While the implementation of Digitalization and Industry 4.0 concepts in the European manufacturing sector is progressing rapidly, there are other aspects that are relevant to achieve the goal of optimal process control. One of them is the "cognitive aspect" that enables process plants to learn from historical data and autonomously react to changes in the running process or even predict undesired events in the process behaviour before they can occur and negatively influence the production. The aim of the project is to add cognitive capabilities to existing process plants and thus provide solutions for dealing with unpredictable behaviour.

Project objectives

The overall project goal is to enhance the performance of cognitive production plants in the European process industries of aluminium, silicon and steel using hybrid and cognitive Digital Twins. Through the extended concept of the Digital Twin, non-optimal behaviour in production processes is to be recognised and predicted, and production control is to be automatically adapted in order to keep product quality and efficiency as optimal as possible. A reference architecture for cognitive Digital Twins is also being developed, as well as a toolbox that will enable the results and functionality to be transferred to other process plants.