From Explosions to Equations

"Prepare for a fast-paced journey through the extraordinary world of explosives, where dramatic history meets the very newest frontiers of Chemistry. This lecture explores how explosive reactions have shaped civilisation, inspired scientists, and pushed chemistry to its limits. From the colossal volcanic eruption responsible for the largest explosion ever recorded on Earth, to the controlled brilliance behind some of the most famous historical discoveries, we will uncover how energetic materials work, why they are so powerful and how chemists study them safely today.

With spectacular live demonstrations bringing the science to life, students will experience the reactions that gave Chemistry its reputation for excitement and wonder. The talk concludes at the cutting edge of current research, including the remarkable 2025 synthesis of what may be the most explosive molecule ever created. Combining breathtaking science, real discovery and unforgettable visuals, this is a rare opportunity to witness the raw power, danger and beauty of Chemistry at its most thrilling."

A physics-informed digital twin for aerodynamic flow systems is developed to reconstruct unsteady wind turbine and wind farm wake fields from sparse real-world measurements. The approach combines line-of-sight (LoS) wind speed data from commercially available continuous-wave wind LiDAR with the Navier–Stokes equations through physics-informed neural networks (PINNs). The LiDAR measurement strategy includes both Plan Position Indicator (PPI) and Velocity Azimuth Display (VAD) scans, providing sparse LoS observations in the flow field. The proposed PINN architecture integrates Fourier feature encoding with fully connected neural networks to enhance the representation of multiscale turbulent structures. Model training simultaneously minimizes the mismatch between predicted and measured LoS velocities (supervised loss) and the residuals of the Navier–Stokes equations evaluated at randomly sampled collocation points (unsupervised loss). After training, the digital twin reconstructs the full spatio-temporal flow field by fusing real-time LiDAR observations with physics-based constraints, enabling recovery of unmeasured flow information. The framework is validated on two configurations: a single wind turbine and a large-scale offshore wind farm case study at the C-Power site in the North Sea. Results demonstrate accurate reconstruction of complex wake dynamics. Detailed analysis confirms that key flow features—including wake propagation, streamwise evolution, and spanwise wake meandering—are successfully captured, highlighting the potential of physics-informed digital twins for real-time wind energy applications.

What could a famous casino possibly have to do with quantum field theory? How can studying melting ice teach us more about subatomic particles? What actually is quantum mechanics? In this talk I will show how these things are more connected than you might think, and how recent advances in computation have allowed us to solve significant problems within the simulation of fundamental physics.