Datadriven modelling of nanofibril dynamics in flowing systems

Design, scale-up and optimisation of flow systems for production of nanostructured assemblies is critical if the potential of many nanotechnologies is to be realised. Computer simulations of the flow and material structure will, of course, play an important role in this work.

However, the behaviour of nanoscale building blocks (such as nanofibrils) in flows cannot be simulated correctly without proper constitutive relations. Datadriven modelling is an attractive route to obtain such relations.

This project will use Bayesian optimisation together with transfer learning to obtain consistent models of nanofibril dynamics in flows, where consistent means that flow, nanostructure and rheology all agree with experimental measurements.

The models obtained will later be used in our efforts to assemble high performance biomaterials, and development of processes for manufacturing of such materials.

The project capitalises on a unique combination of methods for velocity, concentration and structure measurements in flowing nanofluids (optical coherence tomography and X-ray scattering). These experimental abilities are combined with numerical simulations.

Capitalising on these capacities, we will develop a modelling methodology where as simple measurements as possible will be used together with high-end computational fluid dynamics will to model, predict and optimize the nanostructure of novel biobased materials with exceptional mechanical, biological, electrical or thermal properties.