Shell has launched a pilot AI-based catalyst health monitoring system at its Singapore refining complex. Utilizing machine learning models and real-time sensor streams from catalytic reactors, the system tracks performance degradation, contamination risk, and conversion efficiency in near real time.

By analyzing temperature gradients, pressure profiles, and feedstock variability, the AI model flags early signs of catalyst fouling or deactivation—potentially enabling interventions weeks earlier than traditional lab testing cycles. This capability not only extends catalyst life but also improves uptime and yield consistency. In tests, preliminary alerts flagged subtle transitions that would have gone unnoticed until a scheduled shutdown.

For refiners, integrating AI into catalyst monitoring is a high-value entry point to process digitalization—requiring modest instrumentation upgrades but offering outsized operational returns. As margins tighten and feedstocks fluctuate, tighter catalyst management becomes a key differentiator.