In steel production, even minor failures in critical components can have severe consequences. Ladle slide gate systems, which play a vital role in continuous casting processes, are subjected to intense conditions, including heat, corrosion, and mechanical wear. Traditional maintenance approaches, often based on experience rather than data, are increasingly challenged in this environment.
To address this, we developed a comprehensive Health Check Platform in close collaboration with RHI Magnesita and CSEM. The platform utilizes artificial intelligence (AI) and digital twins to provide targeted insights into the condition of critical steel production components. But what does this mean in practice?
The platform continuously collects data from a wide range of sensors and production systems along the steel mill’s value chain. Key indicators such as temperature trends, corrosion levels, and mechanical stresses are gathered to create a detailed condition profile of the ladle slide gate systems. These insights are then fed into digital twins—precise virtual representations that simulate each component’s lifecycle in real-time.
Collecting data alone, however, is not enough to reliably determine the condition of key components.
Only by integrating AI to analyze collected data and identify patterns can we detect early signs of wear and potential weaknesses. The AI continuously improves its predictions using accumulated data and expert feedback, significantly enhancing the accuracy of forecasts and the relevance of actionable recommendations.
Through the combined power of digital twins, AI, and expert knowledge, the platform reliably predicts component health and remaining service life. This enables proactive, precise maintenance planning that helps prevent unnecessary downtime and premature part replacements.
But how is this implemented in practice?
Our latest paper, written by Verena Schmidt, Adi Mehmedovic, Till Schöpe, Christoph Netsch, and Fabio d’Isidoro, provides answers to this question. It addresses the core challenges:
How can diverse data sources in a steel mill be seamlessly integrated?
In manufacturing, the quality of AI insights depends on the variety and availability of relevant data sources. Each steel mill is unique, with varying data sources. A data-driven solution must be flexible enough to incorporate sensor data, production data, and human feedback without excessive integration costs.
How can complex maintenance processes be automated and streamlined?
Gathering data is only the first step; transforming it into actionable insights is essential. The Health Check Platform’s open architecture allows it to integrate seamlessly with existing maintenance workflows, supporting effective decision-making and optimizing processes.
How are the strict requirements for data security and infrastructure met?
Data security is a primary concern in the steel industry, where production facilities are typically isolated and highly secured. To meet these needs, the platform operates fully “on edge,” meaning it runs entirely within the secure local network of each steel mill. This adds extra technical requirements but ensures independent and fully secure data processing directly on-site.
Download the full paper now for a closer look at how we tackled these challenges and brought our Health Check Platform from concept to a fully realized solution in an industrial setting.
