BLOGAnnounced at MongoDB.local NYC 2024: A recap of all announcements and updates — Learn more >

Enabling critical process and quality gains for the world’s second-largest steel producer

ArcelorMittal deploys MongoDB Atlas with Microsoft Azure to optimize production and eliminate quality issues across multiple sites

Factory worker image.


Industrial manufacturing


MongoDB Atlas


Content management, Analytics



Delivering strength and security

Steel is one of the greatest man-made materials, and has changed the world since its invention thousands of years ago. From the domestic kitchen to transport, energy generation, construction and heavy industry, it’s a material that delivers strength, conductivity and resilience.

The basic technique of producing steel has remained largely unchanged for centuries; the processes that surround it, however, are constantly evolving. In an industry that depends on high production volumes, even the smallest changes can make a vast difference to quality and output.

The multinational steel manufacturer, headquartered in Luxembourg, was formed in 2006 as a result of the takeover and merger of Arcelor by Mittal Steel. It is the world’s second-largest steel producer, with an annual crude steel production of 88 million tons as of 2022.

ArcelorMittal operates its own iron ore mines, has a critically important R&D operation, and is increasingly focused on sustainability. It has set itself a target to become net zero by 2050, and is developing a highly effective scrap metal processing operation.

“ArcelorMittal operates across the whole industry, from wiring and cables to strips and plates for use in manufacturing and, for example, cladding for buildings,” explains Roger Sommavilla, Project Manager at ArcelorMittal. “We cover every aspect of steel production in all areas.”

Photo of the worker in front of wind turbines.

A quality initiative to ensure customer satisfaction

Steel is an industry where quality really matters, especially in markets such as automotive manufacturing where standards are exacting. Quality assurance is an area in which the company has made substantial investments. ArcelorMittal launched a project to digitize its supply chain and enhance operational performance. A key aspect of this initiative was to improve product quality, reduce defects, and cut associated operating costs.

“Making steel is a complicated process,” says Xavier Marc, Project Manager at ArcelorMittal. “However, customer satisfaction is the most important objective. Products that don't meet their standards may need to be downgraded for other uses. To counteract this, our idea was to check quality indicators at each stage of the process and identify any problems as early as possible.”

Tracking quality across the production chain is a demanding task, compounded by the fact that data is scattered across multiple databases. With a typical 2km steel coil consisting of around 2,000 separate quality checkpoints for one product characteristic, data volumes are also high and rising fast.

The company’s global R&D team developed a new component E-GPQS (Global Product Quality System), a quality control platform that stores data sourced from ArcelorMittal factories worldwide. Its primary aim was to deliver a centralized resource to aggregate and process data from multiple production lines in multiple factories across the globe.

A cloud-based solution clearly made sense for this application; however, ArcelorMittal also needed a database that would provide the capacity and scalability to enable E-GPQS to operate effectively.

“Reading data scattered across numerous databases was highly complicated, so using MongoDB to decompartmentalize this process is a huge gain.”

Xavier Marc, Project Manager, ArcelorMittal


Quick, easy, flexible and scalable

Specifically, ArcelorMittal needed a solution that would meet several key objectives. It needed to aggregate and reconcile from local database sources, process high volumes of data – up to 2.5TB a year – with minimal or no downtime, and store the data for at least three years. It also needed to support and maintain the monitoring needs of E-GPQS and have the flexibility to create new data-related services quickly and easily.

Xavier had seen a small, in-house use case that leveraged MongoDB Community and realized that a larger-scale managed solution running with Microsoft Azure would be ideal for E-GPQS.

“We saw that MongoDB was easy to use and could work with what we call ‘hot data’ that doesn’t come from a data lake,” he explains. “It could also handle the volumes we required, because we would be managing several terabytes at a time. MongoDB gave us the tool to put that into place.”

Also important was ease of management. For a business the size of ArcelorMittal, the team behind E-GPQS was relatively busy, so having a self-scalable and self-maintaining solution was a highly attractive proposition.

“MongoDB consultants ran a study to evaluate the environment, quantify our requirements, and identify potential issues on the performance and database side,” says Xavier. “We started with three production lines and haven't stopped extending it since.”

Illustration that represents aggregation and processing the data from multiple production lines using MongoDB Atlas.

Saving time and improving efficiency

With ArcelorMittal now able to view critical data in one central location, it can gain essential insights into production and tackle complex data-correlation tasks that enable process improvement. And, with around 40 production lines connected worldwide, the results are already becoming clear.

“For people working around quality, it's a huge time saver,” Xavier notes. “Reading data scattered across numerous databases was highly complicated, so using MongoDB to decompartmentalize this process is a huge gain.”

Using MongoDB Atlas, E-GPQS has enabled the development of the Defects Management Tool (DMT), which integrates artificial intelligence and deep learning based on image analysis. ArcelorMittal has also developed a series of new dashboards and associated key performance indicators that enable wider analysis both at plant level and across the wider production chain to enable larger-scale studies.

“It makes sense to not just focus on the production level, and this central database allows me to go beyond that,” Xavier explains.

“Ultimately, the aim of this central tool was to control our entire process and to take analysis-based actions on production lines to improve efficiency and avoid defects,” Xavier concludes. “We’ve made great progress in only delivering products that comply with customer requirements and have certainly avoided some potential disputes there.”

The E-GPQS R&D team photo.

The E-GPQS R&D team (From left to right, Thibault Chevet, Xavier Marc, Roger Sommavilla and Lauris Monticelli) with the Emerging Technology Award won with the Defect Management Tool developed on MongoDB Atlas.

“We saw that MongoDB was easy to use and could work with what we call ‘hot data’ that doesn’t come from a data lake. It could also handle the volumes we required, because we would be managing several terabytes at a time.”

Xavier Marc, Project Manager, ArcelorMittal

What will your story be?

MongoDB will help you find the best solution.