University of Birmingham

From Text to Tech: Shaping the Future of Physics-Based Simulations with AI-Driven Generative Models

Deadline: Open all year round

Project Description

In the contemporary industry, multiphysics and molecular dynamics simulations are not just tools – they are the catalysts for breakthrough innovations. They empower engineers and scientists to analyze complex physical phenomena, optimize designs, and predict real-world performance with unprecedented accuracy. From aerospace engineering to groundbreaking drug discovery, the applications are boundless.

However, the process of setting up these simulations, including creating geometry, computational meshes, and configuring parameters, is often tedious and time-consuming. Our project is set to transform the way scientists and engineers set up their simulations. We aim to develop a generative model that automates the creation of all simulation files from textual input. The model will integrate large language models, such as GPT-3, with physics models to interpret text and generate input files for simulation software like Elmer (, an open-source alternative to COMSOL), gmsh (, an open-source finite-element mesh generator) and LAMMPS ( an open-source molecular dynamics software).

Imagine inputting the text “Produce a simulation of a standard 20-tooth bicycle gear subjected to a torque of 40 Nm, and analyze the stress distribution across the gear teeth during a full rotation,” and having the generative model automatically produce all the necessary files to run the simulation in an open-source code, complete with computational mesh. The generative model will meticulously create the geometry of the gear, set up the mesh, define the material properties, apply the boundary conditions, and even generate the script needed to run the simulation. All the files will be in the appropriate format and organized in a structured manner. As a user, you simply have to execute the generated script, and the simulation will run seamlessly, providing you with detailed results and visualizations of the stress distribution across the gear teeth.


  • Integrate physics models and language models like GPT-3.
  • Collaborate with a multidisciplinary team of experts.
  • Publish research findings in high-impact journals.
  • Present research at conferences and workshops.

Eligibility Requirements


Master’s, MEng or BEng degree in Physics, Computer Science, Engineering, or a related field (with a classification of at least 2:1).

Strong programming skills (Python, MATLAB, or similar).

Experience with machine learning frameworks (TensorFlow, PyTorch, etc.).

Excellent written and verbal communication skills.

Ability to work independently and as part of a team.


Knowledge of physics simulations, molecular dynamics, and modeling.

Familiarity with simulation software such as gmsh, Elmer, and LAMMPS.

Application Process

For enquiries, please contact Dr Bahman Ghiassi ( [email protected] ).


Verified by MonsterInsights