The application of Generative Adversarial Networks (GANs) for the design of grain-based materials (Tarmac)

Applications are invited for this 4 year EngD (PhD) studentship based at the Faculty of Engineering and the School of Computer Science, University of Nottingham, and partnered with Tarmac. We aim to revolutionise the design of grain-based materials (such as aggregates, asphalt, concrete, sintered metals) by combining unsupervised machine learning techniques to generate virtual material geometries with multiphysics discrete elements software, to model their physical behaviour. Asphalt mixture will be the material of choice for this project, because it is the most extended material to build road surfaces.

The successful candidate will investigate the applicability of Generative Adversarial Networks (GANs) to model three-dimensional textures of granular skeletons, their pore structures and intergranular fillers. The study will apply these concepts to bi-dimensional representations of asphalt mixture. By combining three-dimensional asphalt reconstructions with multiphysics software, based on discrete or finite elements, the aim is to be able to predict the durability of every asphalt road material without the need for asphalt manufacturing or expensive laboratory testing.

Due to the nature of the project, the following skills would be desirable in the successful candidate:
* creative problem solving
* Python, TensorFlow, Pytorch, or C++ programming
* general familiarity with 3D image manipulation software (e.g., Inventor. SolidWorks, Meshlab, etc.)

The Centre wishes to recruit an engineering or numerical science graduate with a first or high 2.1 (above 65% grade mark average) class honours degrees. The funding available restricts the project to UK students only.

Applications
If you wish to apply, please send your CV and covering letter to the EPSRC Centre for Doctoral Training in Carbon Capture and Storage and Cleaner Fossil Energy: [Email Address Removed].