Proteus

The company Treelogic directs a European research project called PROTEUS, which belongs to Program H2020 in the field of data analysis, with a budget of more than 3 million euros and in collaboration with ArcelorMittal.

WHAT PROBLEMS ARE YOU LOOKING TO SOLVE OR IMPROVE?

The reduction of faults in the manufacture of steel products by development of a computerized tool able to predict with some probability what will occur in the industrial production process. This system could lead to significant cost savings by enabling the prevention of errors.

WHICH OF THE 4.0 TECHNOLOGIES ADDRESSES THE PROBLEM?

Using Big Data Analytics, combined with strategies from Data Science, it is possible to analyze, in real time, an elevated volume of data from sensors installed in the production line and to combine these with other data stored previously from the same line, in order to detect statistical correlations.

The conclusions drawn from this joint analysis of information can predict what will occur with some probability and thus influence decisions with an elevated economic significance, such as stopping the manufacture of a steel sheet before the resulting material fails to meet the required quality conditions and hence cannot be marketed.

The PROTEUS project utilizes the Big Data platform Apache Flink, created in Europe, which constitutes a strategic move to reduce Europe’s dependence on technology from the USA, thus contributing to the development of the European Big Data industry.

RESULT OF THE PROJECT AND IMPACT ON THE COMPANY/INDUSTRY

The project will advance in the Big Data field by making contributions to the Apache Flink ecosystem with the integration of a data processing strategy able to simultaneously analyze historic information and information generated in real time. Based on this new ability, two innovative libraries for advanced analyses based on Machine Learning and visual interactive analytics will be implemented, which will be integrated into Apache Flink.

ArcelorMittal will apply the results of the research, thus validating their success, in a complex, real environment where they could yield important cost savings.

Treelogic, in turn, will direct and coordinate the project, and will be responsible for conducting research in new visualization and data interaction strategies. In view of the new characteristics of the computer systems in the Big Data domain, Treelogic will approach this task from two standpoints at the same time. On the one hand, it will research and define new ways of visualizing and interacting with information from the standpoint of human-machine interaction. On the other, from the technological standpoint, it will implement these new mechanisms of interactive visualization with a novel software architecture that enables working with large amounts of data and data generated continuously always in real time.