Physics-guided machine learning for saturation control across variable hydraulic conditions in heap leaching

Abstract

A physics-guided machine learning surrogate, conditioned on hydraulic parameters, that predicts hourly saturation profiles in heap leaching and is trained with a mass-balance penalty on thousands of Richards-equation simulations. Used in a daily predictive controller, it matched a tuned PI controller in tracking a 0.70 saturation target while reaching the target without overshoot and using about 4% less irrigation. Physics guidance cut long-horizon prediction error by about 30% compared with data-only training, and the surrogate ran about 70 times faster than the numerical solver, making large-scale saturation control practical in simulation.

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.