Global sensitivity analysis of blue hydrogen production: A comparative study using machine learning

Abstract

Neural network and random forest surrogates of a blue hydrogen process (sorption-enhanced steam methane reforming with chemical-looping combustion) cut computation time by about 99% while keeping predictive accuracy. Combined with global sensitivity analysis, they show that the CaO-to-natural-gas ratio is the dominant driver of carbon capture and hydrogen production, and compare Sobol indices (fast) with SHAP (more interpretable but much slower).

Dr Paulina Quintanilla

Assistant Professor in Process Systems Engineering

Assistant Professor in Process Systems Engineering at UCL, leading the LOOPS research group. Research interests include machine learning, optimisation and control for processes and physical systems.