Machine Learning for Computational Fluid Dynamics (Coding, Machine Learning & Fluid Dynamics)

As machine learning (ML) transforms our society, the next generation of engineering will also be revolutionised by these models. Computational fluid dynamics (CFD) solves equations to model fluid flow, to allow the design of faster cars and planes, optimise green technologies like wind turbines, enable biotechnology and make computer games and films more realistic. This project will look to apply transformers, the basic architecture of large language models (LLMs) like ChatGPT, to predict fluid dynamics in the same way they predict the next word in a sentence. This will be applied to the simplest example of turbulence, the minimal flow unit, incorporating molecular detail as part of a multi-physics simulation. This will be coupled with cutting-edge techniques like physics informed neural networks (PINNs) and super resolution from generative adversarial network (GANS) all run on a supercomputer with GPUs. During the project, the student will become an expert in machine learning, fluid dynamics and multi-physics simulation, while researching at the forefront of this exciting new field.