Norex is a family-owned company founded in 1983, dedicated to maximizing steel utilization and pursuing 100% recyclability. Norex provides a wide range of services for the needs of steel mills. In this resource-efficient environment, the Process Heartbeat™ Proof of Concept (PoC) demonstrated how existing automation data can be transformed into valuable information for equipment condition monitoring.
Towards Disruption-Free Production Through the Power of Data
The objective of the trial was to determine how the operation of Norex’s industrial crusher could be monitored using artificial intelligence by utilizing data directly from the automation system. A common challenge in traditional maintenance is that faults are often detected only after they have already caused production downtime, which is a critical obstacle to Norex’s goal of optimizing resources.
This PoC project focused on analyzing crusher operation and identifying anomalies using historical data. The solution utilized a machine learning model that was trained to recognize the plant’s “normal state” based on production data from January, for example. Once the model had learned normal operating behavior, it was able to compare it with new data and identify even the smallest deviations that would go unnoticed by humans or traditional alarm limits.
Artificial Intelligence Detects Emerging Problems Before Traditional Methods
During the trial, Process Heartbeat® demonstrated its capabilities in practice: the system began identifying abnormal behavior in data related to the rotor bearings of Norex’s crusher. It detected subtle changes indicating the onset of a mechanical problem.
Subsequent detailed investigations later confirmed the AI’s observation to be correct: an actual bearing fault was developing in the rotor. However, the most significant finding was that the system was able to identify the fault condition considerably earlier than the other monitoring methods in use. In real-time operation, this information would have enabled maintenance activities to be planned in advance, with avoided disruptions and potential production downtime resulting in significant cost savings.
From Reactive Repairs to Data-Driven Prediction
Although this trial was conducted on crusher monitoring, the Process Heartbeat™ solution is completely equipment-independent. The same AI-based model can be applied to monitor any critical part of a process, such as pumps, conveyors, or motors.
In addition to fault prediction, the system provides significant benefits for production optimization and management. Using data, it was possible to analyze whether production efficiency was at the correct benchmark level: for example, the analysis could identify differences between shifts in feed rates and machine loading. The AI revealed differences in how equipment was operated, enabling the sharing of best practices and the achievement of more consistent production. Furthermore, the system can ensure, for example, precise preheating of machinery, which saves energy and reduces wear.
The trial demonstrated that the transition from a reactive operating model to a predictive and data-driven approach can be implemented with a low barrier to entry. Through the operating model we have developed, our customers can easily test the effectiveness of the solution using their own data, allowing the added value and payback period to be verified concretely even before broader deployment.
