Integrated modelling for sustainable and optimized steel manufacturing processes (ProcTwin)

Purpose and goal:

The goal of the ProcTwin project proposal is to increase energy efficiency and product quality in the production of steel in Europe. This is to be achieved by developing an operator support that is based on a self-learning distributed machine learning model and manages entire process chains as a production unit. The model will balance existing process data in interconnected production steps together with digital twins and innovative sensors to explain the impact of local variations.

Expected results and effects:

The expected effects of ProcTwin are to increase the steelmaking industry´s energy efficiency by 5% and reduce carbon dioxide emissions by 3% through reduced scrapping and re-circulation in production. The process tool must be able to be used to detect local deviations and propose measures to achieve the desired product quality, as well as be used as a basis for removing steel blanks from production if there is an indication that the quality will not be achieved and cannot be compensated.

Approach and implementation:

The project consortium will implememt innovative sensors and create digital twins of several interconnected process steps at two steel manufacturing units: SSAB special steel Oxelösund and Global Steel Wire (Celsa) in Spain, to predict material properties based on a natural spread of process parameters. This synthetic data together with process data will be combined in a self-learning distributed machine learning model that will be implemented in a need-based operator support.