10:00 AM, 28 April 2015 to 5:00 PM, 28 April 2015 at 3M Buckley Innovation Centre University of Huddersfield Queensgate, Huddersfield, West Yorkshire, HD1 3DH
Chairs:
Colin Dennis, RSSB.
Coen van Gulijk, University of Huddersfield.
Venue:
3M Buckley Innovation Centre
University of Huddersfield
Queensgate, Huddersfield, West Yorkshire, HD1 3DH
Sponsors:
RSSB & University of Huddersfield
Introduction
The Institute of Railway Research and RSSB have started working on an exciting new approach to risk analysis for the railway industry: BDRA. BDRA or Big Data
Risk Analysis is founded on the intensified use of data in the railway industry. Cheap computer power and the availability of vast amounts of safety-relevant data provide the fertile basis for this development. Though software applications for data analytics are available today, an integrated approach to safety and risk assessment of the GB railways as a whole, based on data-analytics, is still some way off. In this symposium scientists and industry partners discussed the foundations of BDRA. Speakers and visitors were invited to share their views about the potential of BDRA for the future.
Objective
The objective of the symposium was to discuss the foundations of Big Data Risk
Analysis as a novel tool for risk analysis in the GB railway industry.
BDRA
BDRA should support effective and efficient decision-making like any other safety tool. The difference with traditional risk analysis technique is that it depends on a set of big-data infrastructures. This makes BDRA fundamentally different from traditional risk analysis. Instead of constructing a pre-determined structure for risk calculations, such as a fault tree, different sources of safety - relevant information are crawled through to find answers to specific safety questions or to identify new trends and threats in safety.
To date, two specific innovations of the Institute of Railway Research have demonstrated that there is added value to the Big Data approach for risk analysis; RAATS and Close Call. Analysis Software demonstrate that it is possible to gain meaningful safety information from live data feeds of railway signals and unlock sensible safety lessons from unstructured text data. These precedents, and software packages developed by commercial software vendors, support our confidence that BDRA is viable. However, an integrated BDRA system that analyses ALL relevant data is still quite a long way off. Current software tools have to be evaluated, new data sources have to be tapped into and the requirements of future safety and risk systems have to be considered. This symposium opened the discussion for the development of an integrated system by discussing the foundations of BDRA.
Proceedings
Since the objective is to lay down the foundations of BDRA, all contributors have been asked to contribute a paper to the proceedings.
Programme
9:30 Walk in and coffee |
|
10:00 Welcome |
Colin Dennis (RSSB) |
10:10 Defining BDRA |
Coen van Gulijk (HUD) |
10:45 Novelty of BDRA? |
Paul Swuste (TU Delft) |
11:30 break |
|
11:45 Estimating the Approach Frequency to Red Signals – A Big-Data approach |
Julian Stow (HUD) |
Peter Hughes (HUD) |
|
12:45 Lunch |
|
Marcus Dacre (RSSB) |
|
Violeta Holmes (HUD) |
|
14:30 break |
|
Ben Ale (TU Delft) |
|
15:45 Discussion panel |
Colin Dennis (RSSB) |
16:30 Closure |
Coen van Gulijk (HUD) |
16:30 Drinks |