Mathematics PhD Projects

Supervisor: Dr Adil Mughal

Recent simulations that seek to identify the densest packing of monodisperse hard spheres in a cylindrical tube have revealed a remarkable sequence of structures. Such examples of packing problems have a long and fascinating history and their study continues to be of interest to modern science. While the dense packing of spheres in cylinders is relatively well understood, far less is known about the packing of non-spherical objects. In this project you will investigate the dense ordered packing of hard ellipsoids in cylinders. You will use a combination of numerical techniques (simulated annealing, conjugate-gradient methods, parallel code) and analytical methods to investigate structures generated by packing ellipsoids in narrow channels, and compare them against known sphere-packing results.

Supervisor: Dr I Tudur Davies

The interplay between solid particles and thin liquid films is important in industrial processes such as froth flotation, used in mineral separation and in recycling of plastics, in which the films of a flowing foam are used to separate particles based on their surface wettability. Furthermore, liquid foams' peculiar properties and self-assembling nature mean that they continue to find new applications in new  technologies, e.g., for liquid filtering and/or controlled transportation of solid particles in microfluidics. In these processes, it is important to understand the fundamental interaction that occurs between a solid particle and a thin film or a liquid interface. The project offers several possible directions for exploration:

1. Light particles can become adsorbed at a liquid interface or a soap film when capillary forces counterbalance gravitational forces. How does the likelihood of particle adsorption change with the particle’s shape, orientation, and surface wetness? Under what conditions do particles assemble into patterns at the interface, and how is this affected by the curvature of the interface? These questions could be tackled using a combination of theoretical analysis and numerical simulations.
2. When a particle is too heavy to be supported by a film, it descends through it, stretching the film into a catenoid-like neck before detaching and allowing the film to heal. An important question to explore is under what conditions a particle detaches from the film. To address this question, a mathematical model needs to capture the dynamics of the interaction between the falling particle and the elastic film.

Supervisors: Professor Simon Cox,  Dr I Tudur Davies

Aqueous foams, consisting of gas bubbles in a liquid, are familiar from everyday experience, but they are also used industrially. Examples include the flotation process for the separation of metal ores from rock and enhanced oil recovery from porous rocks. Research in this area is directed towards modelling the dynamic properties of a foam to optimize their use in applications.

Possible projects in this area include:

1. Over time, gas diffuses between bubbles in a foam due to the differences in pressure between them. How does the liquid content of the foam affect this process, or the composition of thegas, and how quickly does a bubble with given geometrical properties grow or shrink? Such questions can be tackled with a mixture of theoretical modelling and numerical simulation.
2. Foams have a yield stress, so that they deform elastically below a certain applied stress, and only flow above it. Numerical solutions of a continuum model could elucidate the interplay between elasticity, plasticity and viscosity. Or, on the bubble scale, numerical simulations could predict, for example, the motion and deformation of bubbles in discrete microfluidics, or the effect on foam flow of a continuous phase that itself has a yield stress.

See http://users.aber.ac.uk/sxc/foam.html for further details.

Supervisor: Professor Simon Cox

A dry aqueous foam is a collection of polyhedral bubbles surrounded by thin films. Its high surface area makes it valuable in many industrial and domestic processes, and may even be relevant to developmental biology. Among the many possible local minima of the surface area, subject to the constraint of fixed bubble volumes, we seek ways of determining the global minimum. This could be for a given number of bubbles within a fixed container, such as a sphere or a cylinder, or for clusters of bubbles with a given variation in bubble volumes, or for bubbles in spaces with non-uniform density. Research projects would include a combination of geometrical calculations and optimisation techniques, possibly with numerical and/or machine learning methods.

See http://users.aber.ac.uk/sxc/two_d_clusters.html for further details.

Supervisor: Dr Rolf Gohm

In recent decades there has been considerable interaction between the applied discipline of classical control theory and some pure mathematics in the theory of operators on Hilbert space. For example, the concepts of controllability and observability have a counterpart in the theory of functional models for contractions on Hilbert space and the problem of robust control is related to operator interpolation problems. It is fascinating to revisit such connections in the emerging new discipline of control for quantum systems. The fact that in such systems there are observables which do not commute with each other gives rise to new types of questions which have no counterpart in the classical theory.

In our Quantum Control Research Group we are beginning to work on these questions and a carefully chosen aspect of this topic matching with the previous knowledge of the candidate could be the starting point of a PhD project. A solid basis in functional analysis combined with interest in control theory and quantum physics would be fine.

Supervisor: Dr Jukka Kiukas

This project will develop the theory of noisy quantum measurement processes, focusing on aspects of (in)compatibility between the noise and measurements. Incompatibility - the existence of measurements and transformations which cannot be realised simultaneously on a physical system - is motivated by the foundations of quantum theory, as well as the current development of practical quantum devices. While the foundations reach back to Heisenberg’s Uncertainty Principle, the current focus is on modelling quantum phenomena in the presence of noise. On the one hand, incompatibility between measurements is one such phenomenon, crucial e.g. for cryptography and state discrimination. On the other hand, incompatibility between the process transformations and available measurements limits the implementation. The two aspects are intricately linked, both reflecting the quantum character of measurements. The overall aim is to derive analytical conditions for incompatibility for processes motivated by open quantum systems. Analytical methods will be supported by numerical convex optimisation. One starting point is pure decoherence, for which the problem has proved tractable even for certain large systems. The topic has a broad intersection with active research areas on quantum contextuality, resource theories, correlations, steering, and monitoring of open systems, providing plenty of challenging directions to explore.

Supervisor: Dr Rob Douglas

This project is concerned with innovative modelling of physical processes using optimal mass transfer, and proving rigorous results for the new classes of cost functions introduced. The archetypal optimal mass transfer problem is to find the optimal way to transfer mass from a set U to a set V, amongst a set S of allowable strategies, where optimality is measured against a non-negative cost function c = c(x, y). One interprets c(x, y) as being the cost per unit mass of transporting material from x ∈ U to y ∈ V . The set S is described by a set of measure-preserving mappings. Depending on the cost function, there may be one, many, or no solutions; since the 1980s considerable progress has been made classifying optimisers for a variety of cost functions and settings (e.g. Riemannian manifolds).

Previous studies have been motivated by problems in meteorology, for example nding more sophisticated ways of measuring weather forecast error (and more generally quantities where displacement is a major cause of difference), and aerodynamic resistance, where one is optimising the shape of an object.

Supervisor: Dr Rob Douglas

The aim of this project is to demonstrate nonlinear stability of solutions to the semigeostrophic equations, a model for weather fronts forming and moving. Stability is understood in the sense that if initial conditions are suffciently close in some L^p norm, the solutions remain within a specified distance for subsequent time.

Planar ideal fluid flow in a bounded domain has a vorticity/stream function formulation as a coupled Poisson equation/transport equation. Vorticity is conserved following the flow; possible vorticities must lie in a set of rearrangements. There is a principle, which has its origins in the work of Kelvin, that energy maximising flows should be stable. G.R. Burton has demonstrated nonlinear stability of flows corresponding to strict energy maximisers over the set of rearrangements of a prescribed vorticity. The semigeostrophic equations have a dual formulation as a coupled Monge-Ampere problem/transport equation. There is a Lagrangian conserved quantity, a geostrophic energy, and the Cullen-Norbury-Purser minimum energy principle. The aim is to identify an appropriate existence theory for the semigeostrophic system in dual form, and show that solutions corresponding to energy minimisers relative to the weak closure of the set of rearrangements of a prescribed function are nonlinearly stable.

Supervisor: Professor Gennady Mishuris

Lattice structures are important media possessing properties which are unusual for a continuum. Recently various lattice structures have been used in industrial applications, for example as filters for elastic waves. Optimisation of the geometrical, mechanical and other physical parameters allows also to create specific band gaps in the structure capturing some frequency regimes and to improve its resistance to possible damage/fracture (or reorient it in a less dangerous direction).
The project will deal with an infinite inhomogeneous lattice structure consisting of particles of different masses linked by small rods exhibiting specific mechanical properties. The aim of the research will be to investigate how structural interfaces and faults (cracks) appearing in the lattice structure affect its strength and dynamical properties. Both theoretical and numerical approaches will be used in the analysis. A strong background in Applied Mathematics or Computational Sciences and Engineering would be required.

Supervisor: Professor Gennady Mishuris

Hydraulic fracturing is one of the major techniques of reservoir stimulation employed by the petroleum and gas industry. The essence of hydrofracturing is pumping fluid into a wellbore at a pressure exceeding fracturing pressure of the target formation. As a result, more of the reservoir is reached and more material can be extracted. An important feature of hydraulic fracturing is insufficient, hard to obtain, uncertain and often unavailable geometrical and mechanical data on the in-situ structure, state and properties of rocks and on the changes induced by fracturing. On one hand this decreases the reliability of the input data needed for numerical simulation; on the other hand, it complicates validation of the output results obtained in numerical simulations. Therefore, there is a need to improve mathematical modelling and numerical simulations of the corresponding coupled problem of fracture propagation and fluid flow. Both more theoretically or more numerically oriented projects are available. A strong background in Applied Mathematics or Computational Sciences and Engineering would be required.

Supervisor: Professor Gennady Mishuris

Piezoelectric composites have had various industrial applications. For example, piezo-actuator is widely used in motor-engines. Such devices are composite structures consisting of anisotropic elastic ceramic layers with integrated electrodes. In the process of exploitation, damage accumulated in the structure may lead to it fracturing and can seriously affect the properties of the structure. The respective mathematical problem is formulated as so-called coupled BVPs (boundary value problems) with mixed (partly nonlinear) boundary and transmission conditions. Moreover, the transmission conditions may well represent a so-called perfect or imperfect interface model depending on the material parameters. Finally, singularities of the elastic and electrical fields at the electrode tips are an important factor in the analysis, as they allow prediction of the long-time behaviour of the structure under consideration. A small parameter in the problem (the characteristic thickness of the electrodes) allows the effective use of asymptotic techniques to find an approximate solution. The project deals with analytical and numerical analysis of the problem. A strong background in Applied Mathematics or Computational Sciences and Engineering would be required.

Supervisors: Professor Gennady Mishuris, Dr Adam Vellender

A thin interphase between different materials is a common phenomenon. It may exhibit various properties depending on the physical features of the interphase (it may be a glue connecting two solid bodies or liquid flow between two surfaces or even a complex structure  consisting of elements of a lower dimension (rods, shells) linking two solid bodies). Such interphases bring essentially different properties to the whole structure in comparison with a material without interphases. The challenge is to determine the properties of the new structure, whose complexity arises not only from its geometry but also from the different physical properties of the interphase material. The well-known “phenomenological” engineering approach consists of determining some properties of the structure from experiments and then replacing the interphase with a zero thickness interface between the materials with specific (phenomenologically prescribed) properties. Such an imperfect interface is defined by the respective transmission conditions matching the solution from different regions. Another approach is to evaluate the transmission conditions from the original problem by means of asymptotic analysis, taking into account small parameters and other available information about the interphase. As a result, one can not only obtain accurately justified transmission conditions but also to estimate an error introduced in the original problem by replacing the thin interphase with the corresponding imperfect interface. Projects theoretically or numerically oriented are available and will deal with different mechanical and physical phenomena (multiphysics) appearing within the interphases. A strong background in Applied Mathematics or Computational Sciences and Engineering would be required.

Supervisor: Dr Kim Kenobi

The Cardigan Bay Marine Wildlife Centre has been gathering an observational data set on the bottlenose dolphins, harbour porpoises and Atlantic Grey Seals of Cardigan Bay since 2005.  Most days, a volunteer from the centre goes out on one of the dolphin-watching boats that leave from Newquay harbour in West Wales. There are three main sub-projects within this PhD proposal:

1. We will use statistical modelling to assess the extent to which the likelihood of seeing marine mammals is related to the environmental and human-impact variables.  The modelling will build on generalised linear models (GLMs) (using both logistic (binary) and count (Poisson or negative binomial) models), generalised additive models (GAMs) (GLMs with additional non-parametric smoothing functions included as extra explanatory variables) and generalised estimating equations (GEEs) (an extension of GLMs that incorporates the spatial and temporal correlations in the sightings data).
2. There are other sources of data that need to be incorporated to extend the research to make inference about the social structure of the marine mammals, the calving rates of females and the occurrence of various types of markings and pigmentations/skin lesions.  These data include photo identification data and sound data from hydrophones.
3. As part of the PhD program, we will build a web-based interactive tool that will be on permanent display in the Cardigan Bay Marine Wildlife Centre. It will enable tourists and members of staff of the wildlife centre to visualise changes in distribution of marine mammals since 2005, as well as to look at the results of the models described in point 1. The staff of the Cardigan Bay Marine Wildlife Centre will need training on how to use this tool, and the PhD student would be heavily involved in successfully embedding the technology in the centre.