RailSAFT aims to develop an affordable and reliable Non-Destructive Testing (NDT), automated ultrasonic inspection technique for high manganese, wear-resistant steel rail crossover points (Frogs). These are commonly used on the UK and global rail networks and are susceptible to in-service cracking due to high impact loads from rolling stock. The early detection of cracks at safety critical locations in rail is vital because they can propagate in service and may ultimately lead to failure with potentially catastrophic consequences. Flaws detected at an early stage in their growth cycle can be monitored/ assessed and repaired before risk of failure. Modelling & simulation methods will be used to develop algorithms for the precise control of the ultrasonic beam generated by phased array probes that are to be developed. Synthetic Aperture Focusing (SAFT) together with advanced signal processing will enhance Signal Noise Ratios thus improving defect detection in cast Frog rail sections.

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

There are approximately 70,000 masonry arch bridge spans on the UK road and rail networks (approx. 1 million spans worldwide), the vast majority of which are now well beyond the 120 year life usually expected of bridges. Though masonry arch bridges are in general considered long-lived structures, large numbers are now showing signs of distress. However, the cost of replacing these bridges in the UK alone would run into tens of billions of pounds, and their aesthetic and heritage value is also significant. Unfortunately the methods currently used to assess their capacity are antiquated and/or over-simplistic, making the task of prioritising renewal or refurbishment schemes extremely difficult (the still widely used MEXE method of assessment, which dates back to the 1940s, has very limited predictive capability and offers little scope for future enhancement). Weathering, continually increasing traffic volumes and factors such as the increased frequency of flood events brought about by climate change (affecting bridges over water) only serve to exacerbate the situation. Furthermore, although the primary focus of recent research has been on prediction of structural failure (the `ultimate limit state'), prediction of the level of service load above which incremental damage occurs (the `permissible limit state') is now a key priority for infrastructure owners, who are under increasing pressure to provide transport networks which are resilient. However, a significant barrier to delivering this using existing tools is that current assessment codes prescribe a fixed ratio between the ultimate and permissible load carrying capacities, which, given the diverse range of bridges in the field, is inappropriate and can lead to highly imprecise bridge assessments, and in turn to major economic implications.The present situation stems from our limited understanding of the 'real-world' behaviour of masonry arch bridges, which typically contain soil fill material surrounding and interacting with the arch barrel when loading is applied, and where both working (cyclic) and ultimate loading regimes are important. Developing an improved understanding of such behaviour is the main focus of this project. To achieve this, highly instrumented soil-arch interaction tests will be undertaken, with low-friction, clear sided, medium and full-scale test chambers and state-of-the-art Particle Image Velocimetry (PIV) techniques used to ensure a comprehensive and high quality experimental data-set is obtained. Test variables will include loading type (quasi-static vs. cyclic), bridge type (road vs. railway), fill material type and the presence or otherwise of near-traffic surface strong / stiff layers. Numerical modelling techniques and novel `system identification' techniques will be employed to ensure the full experimentally obtained data-set is used when validating the models developed. Finally, the ultimate objective is to use the improved understanding obtained to develop more rational assessment tools for use by engineers.

The OCCASION project brings together the University of Southampton's expertise in railway simulation and control (Transportation Research Group) with more generic expertise in operational research (Centre for Operational Research, Management Science and Information Systems). This project will identify and assess innovative approaches to overcoming nodal capacity constraints by examining the scope for technological improvements and operational changes. Although the emphasis is on modelling, it will also cover technological and operational issues. This will include examination of incremental changes, such as improved design of points, changes in signal spacing and overlaps, but also more radical changes including concepts from other modes (e.g. intelligent speed adaptation) and a relaxation of the Rules of the Route/Plan. We will adopt a layered approach by examining nodes of increasing complexity on the South West Main Line before developing a detailed case study of Reading station and its approaches. Our methodology will consist of four main elements. Firstly, we will provide a state of the art review which will examine how nodal capacity problems have been tackled to date in Britain and overseas. We will also examine systematic approaches to innovative problem solving, as proposed by the TRIZ methodology and general systems theory. Second, we will develop a generic meso-level model and simulation tool, based on RailSys, which will determine train routeings and schedules, levels of disruption and reactionary delay and measures of capacity utilisation at nodes. Third, we will develop a micro-level optimisation by applying production scheduling techniques to rail scheduling, and by specifically investigating shifting bottleneck procedures and local search approaches. Fourth, we will integrate the simulation and optimisation models by using a multi-commodity integer programming formulation to examine cost versus service quality trade-offs, using techniques we have previously applied to rail freight. This will be used to determine the most effective technological solutions (including enhancements to signalling, switches and crossings) and operational solutions (including dynamic traffic management). In undertaking this work, we will be assisted by our industrial partners, Arup (operations) and Balfour Beatty Rail (technology). Arup will also use the Legion simulation model to determine the extent that pedestrian movements within the station may constrain the scheduling of trains through the station. Our key outputs will be prototype software tools that will assess the extent to which nodal capacity can be increased. This could be subsequently applied to other bottlenecks on the National Rail network. An advice guide would also be produced on measures to overcome capacity constraints at nodes.

Cloud computing offers the ability to acquire vast, scalable computing resources on-demand. It is revolutionising the way in which data is stored and analysed. The dynamic, scalable approach to analysis offered by cloud computing has become important due to the growth of "big data": the large, often complex, datasets now being created in almost all fields of activity, from healthcare to e-commerce. Unfortunately, due to a lack of expertise, the full potential of cloud computing for extracting knowledge from big data has rarely been achieved outside a few large companies; as a result, many organisations fail to realize their potential to be transformed through extracting more value from the data available to them. UK industry faces a huge skills gap in this area as the demand for big data staff has risen exponentially (912%) over the past five years from 400 advertised vacancies in 2007 to almost 4,000 in 2012 (e-skills UK, Jan 2013). In addition, the demand for big data skills will continue to outpace the demand for standard IT skills, with big data vacancies forecast to increase by around 18% per annum in comparison with 2.5% for IT. Over the next five years this equates to a 92% rise in the demand for big data skills with around 132K new jobs being created in the UK (e-skills UK, Jan 2013). While characteristics such as size, data dependency and the nature of business activity will affect the potential for organisations to realise business benefits from big data, organisations don't have to be big to have big data issues. The problems and benefits are as true for many SMEs as they are for big business which, inevitably broadens and increases the demand for cloud and big data skills. Further, even when security concerns prevent the use of external "public" clouds for certain types of data, organisations are applying the same approaches to their own internal IT resources, using virtualisation to create "private" clouds for data analysis. Addressing these challenges requires expert practitioners who can bridge between the design of scalable algorithms, and the underlying theory in the modelling and analysis of data. It is perhaps not surprising that these skills are in short supply: traditional undergraduate and postgraduate courses produce experts in one or the other of these areas, but not both. We therefore propose to create a multi-disciplinary CDT to fill this significant gap. It will produce multi-disciplinary experts in the mathematics, statistics and computing science of extracting knowledge from big data, with practical experience in exploiting this knowledge to solve problems across a range of application domains. Based on a close collaboration between the School of Computing Science and the School of Mathematics and Statistics at Newcastle University, the CDT will address market requirements and overcome the existing skills barriers. The student intake will be drawn from graduates in computing science, mathematics and statistics. Initial training will provide the core competencies that the students will require, before they collaborate in group projects that teach them to address real research challenges drawn from application domains, before moving on to their individual PhD topic. The PhD topics will be designed to allow the students to focus deeply on a real-world problem the solution of which requires an advance in the underlying computing, maths and statistics. To reinforce this focus, they will spend time on a placement hosted by an industrial or applied academic partner facing that problem. Their PhD research will therefore deepen their knowledge of the field and teach them how to exploit it to solve challenging problems. Working in the new, custom-designed Cloud Innovation Centre, the students will derive continuous benefit from being co-located with researchers, industry experts, and their fellow students; immersing them in a group with a wide range of skills, knowledge and experiences.