Summer research in School of Engineering

Project

Structural integrity assessment of offshore wind turbine structures

Research lead

Professor Ali Mehmanparast

Research objective

Offshore wind energy is central to the UK’s transition towards Net Zero, with rapid expansion of offshore wind capacity planned over the coming decades. As existing offshore wind farms age and new developments move into increasingly challenging offshore environments, understanding the long-term degradation and structural integrity of offshore wind infrastructure is becoming increasingly important. This 7-week summer research project will contribute to the long-term research vision and roadmap of WIND-PROLIFE, a major EPSRC Open Fellowship research programme led by Professor Ali Mehmanparast at University of Birmingham. WIND-PROLIFE aims to develop new scientific understanding, experimental/numerical methodologies and engineering solutions to enable the safe life-extension of offshore wind structures.

The summer project will focus specifically on corrosion-fatigue testing of structural steels used in offshore wind applications, using state-of-the-art experimental facilities within the Civil Engineering Laboratories at the University of Birmingham. The student will gain hands-on experience in specimen preparation, corrosion-fatigue testing, data analysis and interpretation of test results.

Project objectives:

- Review corrosion-fatigue degradation mechanisms relevant to offshore wind turbine support structures.
- Develop and implement an experimental methodology for accelerated corrosion-fatigue testing.
- Conduct fatigue tests on offshore structural steel specimens under selected environmental conditions using advanced laboratory facilities.
- Analyse fatigue life, failure behaviour and experimental trends, and compare the results with available data and established fatigue design approaches.

The project is suitable for undergraduate and MSc students interested in offshore renewable energy, structural integrity, fatigue and fracture, materials engineering, and experimental mechanics. It will provide an opportunity to contribute to a major long-term research programme addressing one of the key engineering challenges associated with the future sustainability and life-extension of offshore wind infrastructure.

For informal enquiries, please email your CV together with a brief statement outlining your research interests to Professor Ali Mehmanparast at: a.mehmanparast@bham.ac.uk.

Pre-requisites

Strong background in one or more of the following topics: structural integrity, fatigue and fracture, materials engineering, experimental mechanics, civil engineering, mechanical engineering.

Project 1

Multi-Agent reinforcement learning in robotic metamaterials

Research lead

Dr Jack Binysh

Research objective

How can mechanical collectives of simple robots learn to navigate their environment without external supervision, just by collecting local information from their neighbours?
In this project, you will take robotic metamaterials – architected lattices woven from sensors and actuators – and challenge them to locomote over rough terrain. You will be given existing code to simulate these materials, and will apply cutting-edge multi-agent reinforcement learning techniques to explore optimal policies for their locomotion, and delve into the physics and mechanics behind these policies.

This project is ideal for an engineer with an interest in mechanics and artificial intelligence as applied to unconventional, next-gen robots. You will have many opportunities to present your work and take part in an ongoing collaboration with our international partners in this work, based in the Netherlands. See BinyshLab.com for some images of the robotic materials you will work with.

Research pre-requisites

The project will involve kinematics simulations using existing in-house codes, combined with application of reinforcement learning packages available in e.g. Python. The project can also involve an experimental component, involving mechatronic fabrication, to suit the student.

Project 2

Learning in active mechanical networks

Research lead

Dr Jack Binysh

Research objective

How can we make structures and robots that adapt to their environment after deployment? One strategy comes from ourselves: the neuronal connections in our brains are not fixed, they are plastic - constantly rewiring in response to our environment. This process is called Hebbian learning.

In engineering, giving mechanical networks plasticity and memory can allow materials to learn from their environment. But because a mechanical structure is purely passive, there are severe limitations on what can be learnt.

In this project, you will implement Hebbian learning protocols in active mechanical lattices built from robotic actuators. You will explore how allowing active couplings, i.e. motor strengths, to remodel in response to different environments, e.g. different terrains, can unlock new learnable functionalities like tuneable wave propagation and locomotion.

This project is ideal for an engineer with an interest in mechanics, and who is motivated to learn about physically-grounded methods in machine intelligence like Hebbian learning.

Research pre-requisites

The project will involve dynamical simulations using existing in-house code combined with application of Hebbian learning techniques, likely written in Python. The project can also involve an experimental component, involving mechatronic fabrication, to suit the student.

Project 3

Totimorphic lattices for reconfigurable structures

Research lead

Dr Jack Binysh

Research objective

Can we build lightweight mechanical structures that reshape continuously, tune their stiffness on demand, and recover when individual components fail?

In this project, you will study Totimorphic lattices - mechanical metamaterials built from neutrally stable unit cells that can move through a continuous family of shapes - and explore the potential of these lattices for reconfigurable structures, with applications in e.g. Space.

Depending on your interests, you will either (i) develop computational models or (ii) work with 3D-printed structures and programmable linear actuators to guide a lattice through shape-changing pathways and lock it into load-bearing configurations.

Using tools from nonlinear mechanics and inverse design, you will explore how manufacturing imperfections, friction and actuator failure affect performance. A central challenge will be to devise strategies that allow the lattice to reconfigure around damaged or disabled components, preserving a target shape or mechanical response.

Research pre-requisites

The project will involve basic kinematics simulations, combined with application of inverse design tools, likely written in Python. For an experimentally minded student, the project will involve some 3D printing and mechatronic fabrication.

Project 4

Controlling robotic materials

Research lead

Dr Jack Binysh

Research objective

Nature builds functional organisms using a mixture of bottom-up collective phenomena and top-down control. Why can’t we?

In this project, you will focus on applying control theory to simple artificial “organisms” – collections of robots coupled together mechanically. By themselves, these collectives display a rich variety of patterns. Harnessing and controlling these patterns would open the door to multifunctional swarm robotic systems that are extremely robust to damage. Yet we lack the right control strategies to do so. You will develop these strategies, focusing first on controlling simple two-unit systems to display a variety of different dynamical states, before moving to larger collectives.

This project is ideal for an engineer with a strong interest in control theory and nonlinear dynamics with applications to swarm robotics. You will have many opportunities to present your work and collaborate with our international partners in the Netherlands and Australia.

Research pre-requisites

The project will involve basic mechanical simulations performed with Newtonian dynamics codes that we have, alongside application of concepts from control theory and mechatronics.

Project 5

Edge waves in nonreciprocal mechanical lattices for one-way waveguides

Research lead

Dr Jack Binysh

Research objective

Reciprocity is a basic symmetry in mechanics, which says that signals propagate in an ordinary material bi-directionally: If I can send a signal (e.g. an elastic wave) to you through a material, you can send one to me.

Active materials, for example materials made from robotic constituents, can break this fundamental physical symmetry: a non-reciprocal material will only guide waves in one direction, acting as a continuous mechanical “diode” in which information travels unidirectionally. This one-way behaviour opens many applications in wave-guiding and mechanical computing.

In this project, you will explore how waves at the edge of a non-reciprocal material behave: one-way information transmission implies that these edge waves can be steered around corners and obstacles without backscattering, making an ideal waveguide. You will design waveguides built from combinations of these non-reciprocal building blocks, and characterize how the design of the mechanical lattice impacts their wave-steering abilities.

This project is ideal for an engineer with a strong interest in mechanics, and in particular mechanical metamaterials – architected mechanical lattices with properties (like one-way waves) not found in conventional materials.

Research pre-requisites

The project will involve basic mechanical simulations performed with in-house Newtonian dynamics codes provided to the student, alongside analysis of the equations of motion for these edge waves.

Project 6

Beams that snap by themselves for bio-inspired soft robots

Research lead

Dr Jack Binysh

Research objective

A conventional robotic arm is centrally controlled – each actuator is individually told what angle to take, or what position to move to. By contrast, many natural limbs – from the flagella of a sperm cell to the legs of a starfish – will snap and flex autonomously without a centralized controller. Understanding how to reproduce such self-organized snapping motions in artificial systems could be key to designing small-scale robots, where conventional control schemes are not available.

A simple platform to understand and exploit this self-organised snapping is an active beam built from motorised linkages – see for example our work “Non-reciprocal Buckling Makes Active Filaments Polyfunctional”.

We have developed simple models to understand how these active beams function when snapping only occurs between two modes and in two dimensions. In this project, you will take this understanding to the next level. You will explore either (i) fully three-dimensional self-snapping beams, or (ii) models that can capture multiple different modes of snapping.

This project is ideal for an engineer with an interest in beam mechanics and bioinspired robotics, and will provide great opportunities for collaboration with our partners at UNSW, Australia.