Service

  • Data and AI
  • Strategy and Transformation
  • Technology and Engineering

Case study

AI in Action: Surface Defect Detection in Outokumpu Stainless Steel

Artificial intelligence is much talked about, but solutions running in production are rare. Outokumpu, Europe’s largest stainless steel producer, turned to Nortal, its trusted partner for decades, to create a business-critical AI solution that is set to be scaled globally.

While steel has been manufactured for a thousand years, stainless steel wasn’t created until the early 1900s. And given its complex chemical post-processing procedure, engineers and scientists have never stopped trying to perfect it.

Challenge

Surface defects cannot be avoided even in the most sophisticated production facilities. However, by detecting these defects early in the process, it is possible to better ensure that the customer gets the right quality required for their needs. Whenever a stainless steel surface is on display, such as on a building’s façade, it must be flawless. A car part hidden from view, however, can tolerate some defects. The key is to make sure the quality produced matches the end user’s needs, ensuring both a happy customer and a profitable business.

There is a significant business advantage for manufacturers able to identify surface issues as far upstream as possible, allowing preventive action, reduced cost of reprocessing, and right-quality products routed to customers. Outokumpu, Europe’s largest producer of stainless steel, turned to Nortal to develop an advanced surface inspection system (ASIS) that could be scaled globally to other Outokumpu sites. Nortal is Outokumpu’s partner for its Global Data Platform, enabling seamless collaboration in the scaling processes.

 

Solution

Understanding defects and root causes like never before

“A defect in hot rolling can extend further down to cold rolling, and a scratch can elongate to several meters,” explains Maximilian Hartmann, ASIS project manager from Outokumpu. “Understanding how defects occur helps us improve processes and take action early enough so we’re not doing unnecessary rework.”

Beginning in October 2024, Outokumpu’s Tornio plant will switch to an altogether new version of an ASIS system which uses AI and Machine Learning to identify defects. “Human inspectors can’t evaluate a coil to the same degree of detail and consistency that a machine learning system can,” says Hartmann. “The renewed ASIS system supports inspectors in evaluating the full coil beyond what bare human eye can detect.”

Constructing the new version of ASIS system was especially complicated, since building from scratch was not an option. First, a proof-of-concept ASIS system built by a previous vendor needed to be stabilized. The team utilized as much of the existing solution as possible but were required to re-engineer the application’s hybrid cloud architecture from scratch in order to meet the high availability requirements for the application. As a production-critical solution, ASIS must run even in scenarios when the internet is down, adding to the complexity of the required technology architecture. Nortal’s Senior Data Scientist, Erkki Alamäe, greatly enhanced the AI-based defect detection model, enabling the team to integrate it more tightly into the solution and significantly raise the level of automation in the defect detection process.

Result

Nortal’s solution unlocks massive benefits

“Surface quality is one of the most important parameters of stainless steel production. The new version of the ASIS is a system we can scale globally and use the Machine Learning model to help the operators’ work in production without relying solely on manual labor to ensure right quality for every customer,” says Hartmann.

The benefits are many. Customers get the quality they require, production planning and flow are optimized, and coils with problems are repaired or rerouted. Defects identified earlier mean lower costs, plus long-term benefits thanks to the identification of root causes. It’s a multisite solution that will allow quality information for coils to be transferred between various Outokumpu sites in Europe. Additionally, there are significant savings in storage costs, rejected materials, and through fewer customer claims.

– Maximilian Hartmann, ASIS Project Manager Outokumpu

I’m really happy how the development has gone with Nortal. We always moved constructively in the right direction. The way of working is clean, well-organized, and people talk openly. As the new version of ASIS is taken into use, the full value of ASIS materializes. The value of ASIS will become even more evident as we continue our rollout across additional lines and sites, confirming our belief in its potential.

Related content

Case study

  • Strategy and Transformation
  • Technology and Engineering
  • Energy and Resources

Nortal's cooperation with St1 enables the management of the renewable biofuels value chain

In collaboration, Nortal Finland and St1 have developed the Samba, Rumba, and Maestro applications tailored to manage the value chain of renewable biofuels.

Case study

  • Data and AI
  • Strategy and Transformation
  • Technology and Engineering
  • Consumer
  • Enterprise

Data architecture unlocking new business opportunities for one of Finland’s largest food services company 

Nortal built a data platform that functions as the central ecosystem of Compass Group Finland’s data and intelligence.

Case study

The future of construction safety
  • Data and AI
  • Strategy and Transformation
  • Technology and Engineering
  • Construction

The future of construction safety

Construction has always been one of the most hazardous industries in which to work. Forward-thinking employers across the world are exploring the possibilities of smart personal protective equipment and real-time data analysis to mitigate risk and improve the safety and wellbeing of their employees.

Get in touch

Let us offer you a new perspective.