Creating a Data Platform

There was also a problem with data sources: at daily strategy meetings on furnace use, the various stakeholders (operators, metallurgists, etc.) were relying on different data: the lack of agreement on the data made decision-making difficult.

The data scientists and metallurgists involved in the project collaborated to establish the parameters required to determine the appropriate quality level for each batch, adjust production, and implement monitoring dashboards for production campaigns with numerous indicators and predictive elements. The business value of the web application proof of concept (POC) was confirmed. Cellenza then stepped in to industrialize it.

Key Results

Data Platform Architecture and Foundation

Cellenza spent three months working within one of the subsidiaries managing a furnace to industrialize the POC.

Cellenza’s data experts first set up the data architecture. Before that, the data scientists had been working from .CSV files provided by the metallurgists, compiling all the data from the various sources: Cellenza, therefore, undertook extensive work to trace the data from these files back to their sources to build the “pipes” and mechanics that would allow the necessary data to be obtained daily. Cellenza’s experts determined which Azure technologies to use, helped the subsidiary make the data available, designed the architecture, and set up the data platform foundation.

In a high-security environment, data security must be guaranteed throughout the chain: Cellenza’s experts implemented authentication, monitoring at various stages of data processing, and data quality indicators..

Finally, Cellenza’s experts updated and improved the web application previously built during the POC to ensure that it best met the client’s needs.

The results: 1.7 million euros in annual savings per furnace

In just one quarter, Cellenza implemented a platform with services that can be fully deployed in just 20 minutes.

The results of implementing this data platform were quickly visible. A year later, the subsidiary reported savings of 1.7 million euros as a result of optimizing the furnace production. A few months later, Cellenza was asked to implement this solution at another subsidiary, building on the success of the first project. The solution has since been deployed on three furnaces, allowing the group to optimize alloy production while saving more than 5 million euros per year.

The results for corporate users are also impressive. The various stakeholders now have access to the same data, which is updated daily and contains clear predictive elements. There is now agreement on the data, and decision-making is streamlined.