Search

Ontology is one of the enablers of the Big Data Risk Analysis project (BDRA). Ontology is the systematic classification of domain knowledge that supports the use of different databases in a meaningful way. This pa-per describes the background of ontologies and use cases within the BDRA research programme. Primarily, they are useful for search engines, text analysis and data-linkage and re-use. Also, the analysis of ontologies offers a shimmer into a new way of working in science itself.

Ontologies have been used in diverse areas such as Knowledge Management (KM), Artificial Intelligence (AI), Natural Language Processing (NLP) and Semantic Web as they allow software applications to integrate, query and reason about concepts and relations within a knowledge domain. For Big Data Risk Analysis (BDRA) in railways, ontologies are a key enabler for obtaining valuable insights into safety from the large amount of data available from the railway. Traditionally, the ontology building has been an entirely manual process that has required a considerable human effort and development time. During the last decade, the in-formation explosion due to the Internet and the need to develop large-scale methods to extract patterns in a systematic way, has given rise the research area of “ontology learning”. Despite recent research efforts, ontol-ogy learning systems are still struggling with extracting terms (words or multiple-words) from text-based data. This manuscript explores the benefits of visual analytics to support the construction of ontologies for a particular part of railway safety management: possessions. In railways, possession operations are the protection arrangements for engineering work that ensure track workers remain separated from moving trains. A network of terms from possession operations standards is represented to extract the concepts of the ontology that enable the safety learning from events related to possession operations.

[EN] New technology brings ever more data to support decision-making for intelligent transport systems. Big Data is no longer a futuristic challenge, it is happening right now: modern railway systems have countless sources of data providing a massive quantity of diverse information on every aspect of operations such as train position and speed, brake applications, passenger numbers, status of the signaling system or reported incidents.The traditional approaches to safety management on the railways have relied on static data sources to populate traditional safety tools such as bow-tie models and fault trees. The Big Data Risk Analysis (BDRA) program for Railways at the University of Huddersfield is investigating how the many Big Data sources from the railway can be combined in a meaningful way to provide a better understanding about the GB railway systems and the environment within which they operate.Moving to BDRA is not simply a matter of scaling-up existing analysis techniques. BDRA has to coordinate and combine a wide range of sources with different types of data and accuracy, and that is not straight-forward. BDRA is structured around three components: data, ontology and visualisation. Each of these components is critical to support the overall framework. This paper describes how these three components are used to get safety knowledge from two data sources by means of ontologies from text documents. This is a part of the ongoing BDRA research that is looking at integrating many large and varied data sources to support railway safety and decision-makers.DOI: http://dx.doi.org/10.4995/CIT2016.2016.1825

Computer scientists believe that the enormous amounts of data in the internet will unchain a management revolution of uncanny proportions. Yet, to date, the potential benefit of this revolution is scantily investigated for safety and risk management. This paper gives a brief overview of a research programme that investigates how the new internet-driven data-revolution could benefit safety and risk management for railway safety in the UK. The paper gives a brief overview the current activities in this programme and infers whether big-data techniques provide a sensible addition to the safety and risk sciences. The overview shows that there is added value for introducing these techniques in the safety and risk domain but serious challenges need to be addressed.

El objetivo general de la tesis es determinar las variables de accidentalidad ferroviaria en las que interviene el factor humano y su valoración respecto los colectivos de riesgo del sistema ferroviario español. Para ello se han analizado los registros de accidentalidad del ferrocarril español, a excepción del Ferrocarril Español de Vía Estrecha (FEVE), para su descomposición en cadenas de sucesos, registrando cada uno de los sucesos en una base de datos estadística diseñada para este fin. Analizando los primeros eventos de la cadena de sucesos de los accidentes se han determinado las principales variables en las que interviene el factor humano, y mediante la base de datos de sucesos elaborada se han realizado los análisis descriptivos necesarios para la obtención de los indicadores de valoración de riesgo para los colectivos de riesgo del sistema ferroviario español.

Safety Science has been developed over time with notable models in the early 20th Century such as Heinrich’s iceberg model and the Swiss cheese model. Common techniques such fault tree and event tree analyses, HAZOP analysis and bow-ties construction are widely used within industry. These techniques are based on the concept that failures of a system can be caused by deviations or individual faults within a system, combinations of latent failures, or even where each part of a complex system is operating within normal bounds but a combined effect creates a hazardous situation.\ud In this era of Big Data, systems are becoming increasingly complex, producing such a large quantity of data related to safety that cannot be meaningfully analysed by humans to make decisions or uncover complex trends that may indicate the presence of hazards. More subtle and automated techniques for mining these data are required to provide a better understanding of our systems and the environment within which they operate, and insights to hazards that may not otherwise be identified. Big Data Risk Analysis (BDRA) is a suite of techniques being researched to identify the use of non-traditional techniques from big data sources to predict safety risk.\ud This paper describes early trials of BDRA that have been conducted on railway signal information and text-based reports of railway safety near misses and the ongoing research that is looking at combining various data sources to uncover obscured trends that cannot be identified by considering each source individually. The paper also discusses how visual analytics may be a key tool in analysing Big Data to support knowledge elicitation and decision-making, as well as providing information in a form that can be readily interpreted by a variety of audiences.

In the Big Data era new data sources are available to get insight from human factors in railways. Close Call System (CSS) is one of the data sources which are being researched in the Big Data Risk Analysis (BDRA) project to extract valuable information for risk management. One of the key challenges of BDRA is the visualisation of a large amount of information into a simple and effective display to risk analysis and making-decisions. In this paper we present the research in converting the free text from Close Call data into a spatial representation of networks of words and perform the text visual analysis in order to identify risk categories. For a small number of Close Call records related to level crossings, trespasses and slips, falls and trips, it was possible to identify the different scenarios. Moreover, the results provide an understanding of how Close Call events are described and how it might influence safety on the railways.

A stop signal passed without authority (aka SPAD) is one of the most serious types of incidents in railways since they potentially cause derailments or collisions. SPADs are complex incidents that have been usually analysed as human factors incidents. Human errors of train drivers such as slips or lapses have been prevalent in SPAD incident investigations. In the big data era, alternatives to the traditional methods can be used to support SPAD analysis of whatever kind. Railway systems produce a huge amount of data from a variety of data sources that can be used to get a better understanding of the factors involved in SPADs. This paper describes a first trial within the Big Data Risk Analysis program (BDRA) in order to combine unstructured data from SMIS/IFCS text records with structured data from of railway signals in order to support the SPAD management.

This paper explores the utilization of On Train Data Recorders (OTDR) in monitoring train safety systems use. Train Protection and Warning System (TPWS), Emergency Bypass Switch (EBS) and the Driver's Reminder Appliance (DRA) were used as examples of train safety systems. Using OTDR data to monitor safety systems has to potential to improve compliance with the Rule Book, especially if data can be collected and analysed in real-time. Identification of the deviation from recommended rules that may have safety implications may be useful in the pre-incident investigation.

Computer scientists are quite clear in their belief that the internet is coming of age. They have a firm belief that the enormous amounts of data floating around in the internet will unchain a management revolution of uncanny proportions. This revolution is referred to as ‘Big Data’. Yet,\ud to date, the potential benefit of this revolution is scantily investigated for safety and risk\ud management of the railways. This work reports about an investigation how Big Data can contribute\ud to safety systems for the GB railways. The experience that is gained also sheds light on Big Data as a driver for change in the railway industry as a whole.