DrGeorgios Rigas

My research group develops Physical AI for complex fluid systems: machine learning and reinforcement learning methods that interact with, learn from, and control real physical processes governed by nonlinear dynamics, turbulence, and partial observations.

Our research sits at the interface of fluid mechanics, control, data-driven modelling, and artificial intelligence. A central theme is to combine physical understanding, high-fidelity simulation, sparse sensing, and optimisation to enable reliable decision-making in realistic engineering environments.

We are particularly interested in:

• Flow control (Reinforcement Learning and Model Predictive Control) for turbulent flows under partial observability
• Physics-informed and data-driven modelling of turbulent and transitional flows
• Digital twins and real-time optimisation for energy, transport, and industrial applications
• High-speed and hypersonic flow physics, including instability, transition, and control
• Experimental and computational fluid mechanics, with an emphasis on methods that bridge simulation and laboratory data

Our goal is to design autonomous learning systems that are sample-efficient, physically grounded, and deployable in real engineering settings. This includes problems where only limited measurements are available, where governing physics is essential, and where models must generalise across operating conditions.

Applications of this work include aerodynamics, wind energy, transport, propulsion, and low-emissions technologies, with the broader aim of supporting the transition to more efficient and sustainable engineering systems.

More details and selected software/projects can be found on the group GitHub: RigasLab.