ABOUT

Project ambition

AI4EAFRoute aims to accelerate the digital transformation of Electric Arc Furnace (EAF) steelmaking by combining advanced sensing technologies, artificial intelligence, and data-driven decision support systems. The project seeks to enhance process efficiency, operational reliability, worker safety, and environmental performance by enabling real-time monitoring, predictive maintenance, and intelligent process control across EAF operations.

By integrating innovative sensors with advanced AI models, AI4EAFRoute will provide steel producers with actionable insights to optimize melting operations, improve electrical performance, reduce unplanned downtime, extend equipment lifetime, and lower energy consumption and CO₂ emissions. The project further aims to demonstrate the scalability and industrial applicability of these technologies, supporting a more competitive, resilient, and sustainable European steel industry aligned with the objectives of the green and digital transitions.

Challenges

The successful implementation of AI4EAFRoute requires addressing several scientific, technological, and industrial challenges associated with the digitalization of Electric Arc Furnace (EAF) steelmaking.

A key challenge is the collection and integration of high-quality data from multiple heterogeneous sources operating under harsh industrial conditions. The project must ensure reliable acquisition, synchronization, and pre-processing of large volumes of sensor and operational data to provide robust inputs for AI model development.

Another major challenge lies in the development of accurate, reliable, and explainable AI models capable of supporting decision-making in complex and dynamic EAF environments. The models must be able to operate effectively despite data variability, changing process conditions, and limited availability of labeled datasets for specific events or failures.

The project must also address the challenge of integrating AI solutions into industrial workflows through user-friendly Decision Support Systems. Ensuring operator trust, usability, and seamless interaction between human expertise and AI-generated recommendations is essential for the successful adoption of the developed technologies.

From an industrial perspective, validating the proposed solutions under real operating conditions represents a critical challenge. The technologies must demonstrate measurable benefits in terms of process efficiency, equipment reliability, maintenance optimization, energy consumption, and CO₂ emissions reduction while maintaining production continuity and safety standards.

Finally, AI4EAFRoute must ensure the scalability, transferability, and long-term exploitation of its results, enabling the developed solutions to be adopted across different EAF plants and operational contexts within the European steel industry.

Work Packages

The overall project is planned to last 48 months and it is divided in 6 Work Packages (WPs).

WP1 focuses on project coordination and management, covering financial, administrative, and technical aspects, as well as societal, ethical, and legal considerations. It also ensures continuous engagement and feedback from the industrial partners throughout the project lifecycle.
WP2 is dedicated to the development, installation, and operation of both innovative and state-of-the-art sensors, as well as the collection of the data required to support the project objectives. This work package therefore addresses the hardware-related components of the project.
WP3 focuses on the development of advanced AI models tailored to the specific objectives of each use case. The activities will begin with comprehensive data preprocessing and integration to ensure high-quality datasets, followed by iterative model development, training, validation, and optimization to achieve robust and reliable performance.
WP4 addresses the integration of the developed AI models into a comprehensive Decision Support System (DSS). Particular attention will be devoted to usability, human-machine interaction, and operator acceptance, exploring innovative approaches to improve decision-making compared with conventional tools, including the potential application of generative AI solutions. Together, WP3 and WP4 constitute the core software development activities of the project.
WP5 is devoted to the demonstration and validation of the developed technologies in representative industrial scenarios. The work will assess the impact of the proposed solutions in terms of CO₂ emissions reduction, energy efficiency improvement, maintenance cost optimization, and enhanced plant reliability and operational performance across the different use cases.
WP6 covers dissemination, communication, and exploitation activities, ensuring wide visibility, stakeholder engagement, and effective uptake of the project results. Building upon the outcomes generated in all preceding work packages, WP6 will maximize the project’s impact and support the long-term exploitation and replication of the developed solutions.