Current Research Projects
Machine Learning for Leidenfrost Droplet Dynamics
Developing operator learning models and Physics-Informed Neural Networks (PINNs) for predictive modeling of multiphase Leidenfrost droplet dynamics. This SNSF-funded project leverages a dataset of 1,000+ high-fidelity CFD simulations to train surrogate models that can predict droplet behavior in real-time.
Viscoplastic Droplet Impact on Leidenfrost Surfaces
Comprehensive numerical analysis of non-Newtonian droplet rebound on heated surfaces using OpenFOAM. The project characterizes interfacial oscillations, dimple formation, and vapour-layer dynamics during impact and rebound of viscoplastic droplets.
iteraSim RAG - Agentic AI for OpenFOAM
Development of an agentic Retrieval-Augmented Generation framework integrating vector databases (Qdrant) with specialized LLMs for the OpenFOAM community. The platform provides version-aware documentation retrieval, automated dictionary generation, and intelligent troubleshooting.
ITER Test Blanket Module Analysis
Numerical investigation of flow, heat transfer, and tritium transport in packed bed breeder blankets for fusion reactors. Development of heterogeneous LTNE-based solvers with radiation effects and multi-region tritium transport models in OpenFOAM.
Packed Bed Scaling Laws & Correlations
Development of validated scaling laws for fluid flow and heat transfer in packed beds filled with monosized and binary pebbles. This work, selected for front cover publication, provides comprehensive correlations for pressure drop and Nusselt number predictions.
Global CFD-ML Workshop Series
Design and delivery of international online workshop series on Computational Fluid Dynamics and Machine Learning integration. Covering fundamentals to advanced topics including PINNs, operator learning, and AI-assisted CFD simulations.