Lecture 1

Date & Time: 11th April 2017, 14:00


Prof Chris Baker (University of Birmingham)                                                                     Dr Hassan Hemida (University of Birmingham)

Title: The context and use of Computational Fluid Dynamics in the study of train aerodynamics

Summary: This presentation will begin by outlining a range of practical issues that need to be addressed in the study of the aerodynamics of trains  – the aerodynamic drag of trains, railway tunnel pressure transients and sonic booms, train induced pressures on trackside structures, stability of trains in high crosswinds, the effect of train slipstreams on passengers and trackside workers,  and the flight of railway ballast under high speed trains. The application of a range of CFD techniques for the study of the last three of these issues (crosswind effects, train slipstreams and ballast flight) will then be discussed and the results of a number of recent projects will be presented, including verification of the CFD results against full scale and model scale experimental data. The presentation will then discuss the wider context of the CFD calculations and how they are applied to address the underlying problems. The main point to emerge is that the CFD calculations are usually just one part of a broader methodology and the accuracy of the calculations and the resources they require need to be appropriate to the overall context.

Lecture 2

Date & Time: 11th April 2017, 15:30

Speakers: Prof Michal Kocvara (University of Birmingham)

Title: New Numerical Tools for Very Large Scale Topology Optimization 

Summary: This presentation starts with a brief introduction on topology optimization as a mathematical tool for optimal design of mechanical components. Although now routinely used in the industry, software for topology optimization suffers from limitations, in particular when used for complex three-dimensional structures. Several  ways will be presented on how to substantially improve efficiency of topology optimization software using modern tools of numerical linear algebra and numerical optimization. These are based on domain decomposition and multigrid techniques and, for the more involved problems, on decomposition of large-scale matrix inequalities using recent results of graph theory.