AFV Beltrame Group: Adopting AI to improve accuracy and performance in the crucial process of furnace loading
A dynamic step in steelmaking efficiency
To optimize plant efficiency in scrap classification and selection, AFV Beltrame Group accelerated enterprise AI adoption. The company built a solution using the generative AI hub feature in the SAP AI Core foundation that accurately checks and grades scrap metal and creates an optimal and efficient mix for steel mill production.
| Industry | Region | Company Size | Partners |
| Mill products | Vicenza, Italy | 2,900+ employees | Regesta Group srl and Clarex srl |
Group CIO, AFV Beltrame Group
Enhancing efficiency and accuracy in quality control processes
Active since 1896, AFV Beltrame Group is a giant in the steel sector in Italy and across the European market, known for manufacturing rolled steel, iron, and bars used in construction, shipbuilding, and excavation. The company operates three steelmaking plants in Italy, one in France, one in Switzerland, and two in Romania.
In an industry that is characterized by energy-intensive processes, meeting the mandatory European Union Emissions Trading System (EU ETS) mechanism is a significant challenge. As such, the company’s business development activity is guided by a commitment to sustainable productivity and conducting business in a way that efficiently and strategically manages the available resources. With a strong focus on sustainability, the organization recycles 97% of its process waste and uses scrap metal as a raw material.
Ready to turn the EU ETS challenge into a strategic opportunity to strengthen competitiveness and accelerate a transition to a more sustainable industrial model, AFV Beltrame Group embarked on an AI adoption journey—one that focused on using AI not only for collaboration but also for helping the company optimize and automate its steel production processes.
Using AI to accurately check and grade scrap metal
In his role managing the IT department across seven European plants, Emanuele Turra, group CIO at AFV Beltrame Group, says that it was important to focus on a strategic AI initiative: “We wanted to develop a solution to support the complex process of scrap classification and selection, as well as to create the most accurate and efficient mix for steel mill production.”
Furthermore, the goal for AFV Beltrame Group was to have a support system that could assist in managing the steel mill operation according to the scrap mix charged on the electric arc furnace.
To bring this vision to life, AFV Beltrame Group partnered with SAP gold partners Clarex srl and Regesta Group srl. The partners demonstrated a deep understanding of steel industry processes and the specific needs of AFV Beltrame Group using AI integrated throughout its enterprise architecture.
Together, the teams developed an AI model to enable semiautomatic identification and classification of scrap material using SAP Business AI capabilities, including the generative AI hub feature in the SAP AI Core foundation. Referencing thousands of labeled images and videos, the AI model analyzes real-time video feeds from cameras monitoring the scrap basket loading process to accurately determine scrap types and their exact positioning in the basket.
These classifications are then directly correlated with electric arc furnace operational data, providing real-time recommendations for adjusting key running parameters such as electricity, methane, oxygen, lime, and carbon to optimize energy consumption and control quality.
Group CIO, AFV Beltrame Group
Improving the precision of steel scrap classification
AFV Beltrame Group and Clarex spent six months calibrating the AI-enabled solution before going live, with early indications providing decisive results. Turra comments, “We’ve been operational for around two months, and we are in an advanced stage of the proof of concept. The results we are seeing are really promising. First of all, we are testing the robustness and scalability of SAP Business AI, which is helping us understand how we can further take advantage of the offerings and their capabilities.”
Awareness of the project within the business is another key success factor. Turra says that the company’s employees are seeing firsthand how AI can become a true work companion. “Even in very complex scenarios, we’ve seen very high accuracy in the scrap and charger recognition under very challenging visual conditions.”
AFV Beltrame Group has also discovered additional uses for the trained AI engine that can assist in classifying scrap when raw materials arrive at the plants.
Improving efficiency, stability, and overall process performance
Based on the initial results, Turra expects to see significant benefits and decreased operating costs from optimizing the electric arc furnace charging process at AFV Beltrame Group. As well as improved efficiency and more-consistent steel quality, the company can expect to reduce its carbon footprint, further enhancing the sustainability of steel production.
Featured partners
Regesta Group srl is an SAP gold partner and a leading name in IT consulting in Italy, providing professional solutions based on the latest generation of software from SAP. Offering deep experience in the mechanical, steel, logistics, and manufacturing sectors, Regesta led the initiative to adopt AI features to enhance efficiency and productivity in AFV Beltrame Group’s steel production processes.
Clarex srl is part of Regesta Group. As an SAP gold partner and value-added reseller, the company offers advanced expertise and tailored support to help companies make the most of their SAP software. Working in tandem with Regesta, Clarex provided proven expertise to guide AFV Beltrame Group’s digital transformation efforts.