Advances in online monitoring and data collection offer opportunities to enhance the efficiency, sustainability, and profitability of engineering processes. Despite the promise of “Big Data” and its impact on other sectors, its application in the chemical and minerals processing industries has not yet reached its full potential.

Our mathematical modelling and machine learning group aims to address this gap by integrating fundamental knowledge with statistical techniques. We focus on using theoretical process models and data-driven machine learning algorithms, often in combination, to improve the operation, monitoring, and control of chemical plants. Applications include fault detection and diagnosis, causality analysis, reinforcement learning, and hybrid modelling, covering fields from minerals processing and petroleum refining to environmental engineering.