Giovanni Grasselli

University of Toronto

ML-FDEM: From advanced research tool to practical engineering solution

The combined Finite-Discrete Element Method (FDEM) has become a powerful numerical approach for simulating complex failure processes in rock and other brittle materials. However, its high computational cost and modelling complexity have limited broader adoption in routine engineering practice. ML-FDEM integrates machine learning with high-fidelity FDEM simulations to bridge the gap between advanced numerical modelling and practical engineering application. By learning from detailed simulation data, ML-based models can reproduce key system responses with substantially reduced computational effort while preserving the essential physical insights provided by FDEM. The framework supports rapid prediction, sensitivity