We are proud to present SAIS LAB’s and Euro-Titan’s team, article published in Recycling Magazine International, Winter 2026, under the title “Metal-from-waste processing in perfect ‘health’.”
The article focuses on the Euro-Titan project and, more specifically, on the role of the AI-powered Big Data Platform, or BDP. The central idea is that titanium production from metallurgical residues is not only a materials-processing challenge, but also a data-management and decision-support challenge.
Euro-Titan works with complex feed materials, such as bauxite residues and titanium dioxide production residues. These materials pass through physical, pyrometallurgical and hydrometallurgical process steps, including hydrogen plasma reduction. During these steps, large volumes of data are produced from sensors, laboratory measurements, inline LIBS monitoring, X-ray diffraction analysis, process parameters, energy use and hydrogen consumption. The difficulty is not simply collecting data, but transforming it into timely, reliable and operationally useful decisions.
The article uses a very effective metaphor: the Euro-Titan process is described as a living patient under continuous medical care. Furnaces, reactors, pumps and pipes are compared to organs, while gases, liquids and slurries are compared to blood flows. Sensors become the equivalent of thermometers, ECGs and blood tests. In this logic, the Big Data Platform acts as the medical intelligence system that continuously monitors the health of the industrial process.
SAIS LAB’s contribution is therefore framed around intelligent process monitoring, data integration and AI-supported optimisation. The platform will integrate heterogeneous data streams into one system, detect early deviations, anticipate equipment or quality anomalies, and support operators with interpretable recommendations before problems lead to downtime. This is particularly important because metallurgical plants generate thousands of signals every second, creating a risk of data overload.
The article also explains two major AI challenges. At the beginning of process development, there is too little operational data to train robust models. Later, during scale-up and live operation, there is too much data to process efficiently. Euro-Titan addresses this by building a curated “teaching dataset” from pilot runs, laboratory experiments, simulations and expert-labelled process knowledge.
Finally, the article highlights the use of CFD models and digital twins. CFD provides a virtual view inside the furnace, while machine-learning models can approximate furnace behaviour fast enough to support real-time optimisation. Together, these tools help reduce energy and hydrogen consumption, stabilise process flows, support preventive maintenance and improve titanium recovery and quality.