Skip to main content

Research Theme: Modelling of Materials and Structures

Research Theme: Modelling of Materials and Structures

We investigate the processing-microstructure-properties-performance relations of materials.

Processing, microstructure, properties and performance

We investigate the processing-microstructure-properties-performance relations of materials. This is essential in the design of advanced engineering materials and components. We use multi-length scale modelling, from atomic to macroscopic level, in the design of materials and components.

We have developed a variety of predictive models.

Diffusion kinetics models predict the necessary processing parameters, such as optimum heat treatment temperature.

Phase-field models predict the microstructure evolution under different thermo-mechanical conditions of engineering materials. These include steels, Ti-alloys, Zr-alloys, Al-alloys, and shape memory materials.

Phenomenological models predict macroscopic properties such as yield strength. They are based on the microstructure predicted by the phase-field models

Phase-field models need data about material properties, such as elastic constants and interfacial properties. Our collaborators provide this data from lower length scale modelling approaches. These approaches include ab initio calculations and molecular dynamics simulations.

Finally, we design engineering components with the predicted microstructure and properties. We simulate performance under different thermo-mechanical conditions using finite element models (FEM). For example, we use ANSYS for some of our simulations.

Structural optimisation

Structural optimisation addresses the complex nature of many engineering structures. They may have tens, or even hundreds, of parameters. Some are continuous, such as the radius of a dome roof or the position of a circumferential girder. Others are discrete, such as the number of radial girders or the type of industrial section used for each girder.

Structural optimisation is multi-objective in nature. There is a need to minimise manufacturing cost: for example, by reducing mass. We must do this without sacrificing structural response and safety. This could include, for example, reducing deflection. These objectives can conflict.

Optimisation methods are evolutionary. They change incrementally from an initial design.

Metaheuristics are problem-independent techniques. Optimisation methods are also metaheuristic. They are often trial and error based.

Many metaheuristic algorithms are nature-inspired, such as the Bees Algorithm.

Optimal solutions are rare. The best algorithms may incorporate aspects of:

  • the engineering knowledge and experience of evolutionary structural methods (such as MESO)
  • the intuition of random searches
Left: MESO step diagram, Right: Mass Deflection Pareto Fronts