Safety performance data analysis

Research Project: Development of a Framework for Incident and Accident Precursor Analysis in Air Transport

Background

To maintain the number of air transport accidents directly related to the provision of Air Traffic Management (ATM) as low as possible, the ATM system is developing ways of improving and monitoring its safety performance level. Air transport Safety Management System (SMS) requires that incident databases be created and maintained in order to provide ways of monitoring safety performance levels.

The search for precursors based on safety related incidents data (where no accident occurred but safety was jeopardised) is becoming more greatly exploited. However, several safety methodologies describing different concepts of accident mechanisms are currently in use, e.g. linear cause and effect models, accident barriers analysis or systemic accident models. Each methodology has an impact on the data collection, processing and analysis leading sometimes to potential bias. Additionally, the quality and reliability of the data collected needs to be fully understood to remove any further bias in analysis results. With database size constantly increasing, both a safety issues identification tool and a decision making process for prioritisation of issues are necessary.

Methodology

This project proposes therefore a framework for a systematic analysis of incident databases supported by a methodology for a robust data quality and reliability assessment.

The analysis framework proposes more sophisticated techniques which integrate investigation results of each incident occurrence to the general model of Air traffic Management functionalities. It identifies which functions of ATM or its sub-components have been affected to better target solutions and recommendations.

The development of this analysis methodology includes the extensive review of aviation taxonomy of standards and definitions essential to the classification of attributes, contributing and explanatory factors used during the incident reporting process. The data quality and reliability assessment tool benchmarks current databases to an optimal template adequately designed for subsequent analysis.

Collaborations

In this particular study, incidents data from several civil aviation authorities and air navigation service providers were processed. Such assessment provides a database owner with a better knowledge of its data content in terms of level of details and representativeness of its system safety level.

Overall, this study aims to provide a consistent framework for the analysis of incident databases which take into account the quality and reliability of the databases.

Publications

This thesis was submitted in January 2012 and the PhD was awarded in April 2012. Two conference papers have been prepared and a number of journal papers are in preparation.

Dupuy, M.D., Majumdar, A. and W.Y. Ochieng (2012) The management of separation-related incidents by air traffic control, Proceedings of the 5th International Conference on Research in Air Transportation - ICRAT 2012. University of Berkeley, California, USA, May 22-25 2012.

Research Project: Modelling Airport Surface Safety

Background

The steady growth in air traffic has been one of the major features in transportation over the last 50 years and forecasts indicate a further growth for at least the next twenty years. In addition to problems associated with congestion and delays, this growth has considerable safety impacts. One major area highlighted by a number of aviation authorities is that of airport surface safety, in particular runway and taxiway safety.

Although previous and current initiatives increasingly emphasize this topic, the industry is characterized by a piecemeal approach and surface safety rarely considered in an integrated manner. To address this issue, this research proposes to develop a model of airport surface safety.

Methodology

As a first step, a theoretical model of normal airport surface operations is developed. Subsequently, a global study of the critical factors that underlie airport surface accidents and incidents (occurrences) is conducted, and used to develop a new holistic taxonomy of causal and contributing factors. The taxonomy incorporates the viewpoint of all relevant aviation stakeholders (regulators, Air Navigation Service Providers, airlines, airport operators,ground handling companies, Accident Investigation Boards) involved in the subject matter. Its robustness is facilitated by the application of different research methods (literature, multi-national safety data analysis, airportssurvey, interviews). In a thirst step, statistical analysis is used to identify the impact of airport characteristics (e.g. airfield geometry, level of equipment, operations) on safety occurrences. Airports can then be categorized in terms of airport surface risk.

The final model of airport surface safety assesses the functional relationship between accidents and incidents and their underlying critical factors in order to outline effective safety mitigation strategies. The model considers the viewpoints of all relevant aviation stakeholders and accounts for data quality issues (i.e. weighting of different databases). All models are validated through accident/incident data and observational data at selected representative US and European airports.

Collaborations / Project Partners

Data was provided by the UK Civil Aviation Authority (UK CAA), Federal Aviation Administration (FAA), New Zealand CAA (NZ CAA), Avinor, OSL Lufthavn, Norwegian A/S, Signature Flight Support and easyJet. The results have been validated with subject matter experts from EUROCONTROL and the FAA.

Publications

1. Wilke, S., Majumdar, A. and W.Y. Ochieng (2012) A Holistic Approach toward Airport Surface Safety, Transportation Research Record: Journal of the Transportation Research Board, Washington D.C., USA, (accepted).
2. Wilke, S. and A. Majumdar (2012) Critical factors underlying airport surface accidents and incidents: A holistic taxonomy, Journal of Airport Management, 6 (2), pp. 170-190.
3. Wilke, S., Majumdar, A., and W.Y. Ochieng (2012), Assessing the quality of aviation safety databases: An external data validation framework, 5th International Conference on Research in Air Transportation — ICRAT 2012, University of California, Berkley, 22-25 May 2012.
4. Wilke, S., Majumdar, A. and W.Y. Ochieng (2012) A holistic approach towards airport surface safety, Transportation Research Board – 91st Annual Meeting. Washington, D.C. USA, Jan 22-26 2012.
5. Wilke, S., Majumdar, A. and W.Y. Ochieng (2011) Analysis of critical factors underlying airport surface safety occurrences – A global comparison, 1st Conference of Transportation Research Group of India (CTRG). Bangalore, India, Dec 7-10 2011.
6. Wilke, S. and Majumdar, A. (2011), A multi-national causal analysis of airport surface safety occurrences, 11th AIAA Aviation, Technology, Integration, Operations (ATIO) Conference. Virginia Beach, USA, Sep 21-22, 2011.
7. Wilke, S., Majumdar, A., and W.Y. Ochieng (2011), The potential of automation to improve airport surface safety, 1st International Conference on Application and Theory of Automation in Command and Control Systems (ATACCS), Barcelona, Spain, 26-27 May 2011.

Research Project: A Framework to support operational Decision-Making in Offshore Helicopter Transportation

Background

In offshore energy exploration and production activities, helicopters play a vital role in the movement of people to installations in the sea. However, this particular activity is also prone to accidents, especially in conditions where visibility is poor, e.g. the nighttime. Given that some of the World’s major energy production facilities are in the seas of countries where the climatic conditions require nighttime flights, e.g. Norway, UK, USA, and Canada, then such operations pose concerns to both regulators and operators. Furthermore, with major energy reserves discovered offshore in emerging nations, especially Brazil and Russia, together with attendant pressure to exploit such reserves with nighttime flights for example, the need to understand the factors that underlie such accidents, with a view to their future avoidance, will become all the greater in future.

Human factors have already been pointed as the main driver to safety on such operations. Less attention, however, has been paid to understanding such factors under a systemic approach which accounts for all their drivers. To date, a number of issues relevant to safety during offshore flights in degraded visual environments (DVE), especially at night, have been identified and remain unaddressed. Gaps exist, for example, in the absence of an integral approach to crew training, interaction between crew members, crew readiness, human-aircraft interface and offshore installation characteristics, environmental conditions, operating procedures, organisational support and safety oversight. As a consequence, helicopter crews consistently make wrong decisions and high accident rates result.

Research Objectives

This thesis aims to develop a framework to support operational decision-making in offshore helicopter operations in DVE, based upon the factors that contribute to increased risk scenarios. The framework should support informed mission acceptance or denial in the strategic and tactical stages, respectively to be used by regulatory bodies, oil & gas companies and helicopter service providers, as well as pilots. Six research objectives have been formulated to achieve this aim:

i) Develop a systemic approach to offshore helicopter transportation, identifying relevant actors and the functional relationships between them. This involves Hierarchical Task Analysis (HTA) and Functional Resonance Analysis Method (FRAM) for example.

1. Explore the extent to which flying in DVE consists a problem to offshore helicopter transportation. This involves multivariate incident and accident analysis using mixed datasets. Data quality considerations are highly important at this stage.
2. Identify the contextual factors relevant to safety in offshore helicopter operations in DVE. This involves in-depth interviews with active pilots in selected scenarios worldwide. Cognitive Task Analysis is used, followed by the application of Grounded Theory, Template and Content Analyses to generate theory from routine, often unreported, experiences with risk factors.
3. Model a context indicator which incorporates the interactions between contextual factors in a probabilistic framework. This involves mathematical modelling applied to Human Reliability Analysis Techniques (e.g., Cognitive Reliability Analysis Method - CREAM and FRAM).
4. Test the usability and validity of the context indicator. This involves observations and experiments with pilots in carefully selected flight simulators around the world.
5. Recommend strategies to enhance resilience in offshore helicopter operations in DVE. This involves interviews with high profile stakeholders in the worldwide offshore helicopter industry.

Collaborations

This research is conducted in conjunction with numerous offshore helicopter operators, regulatory authorities and the EASA.

Publications

1. Nascimento FAC, Majumdar A, Ochieng WY (2012) A multistage multinational triangulation approach to hazard identification in nighttime offshore helicopter operations, Reliability Engineering & System Safety, Vol. 108, Pages 142-153, ISSN:0951-8320 (doi)
2. Nascimento, F.A.C., Majumdar, A. and S. Jarvis (2012) Night-time approaches to offshore installations in Brazil: Safety shortcomings experienced by helicopter pilots, Accident Analysis & Prevention, Volume 47, July 2012, Pages 64-74.
3. Nascimento, F.A.C., Jarvis, S and A. Majumdar (2011) Factors Affecting Safety During Night Visual Approach Segments for Offshore Helicopters, The Aeronautical Journal of the Royal Aeronautical Society.
4. Nascimento, F.A.C., Majumdar, A., Ochieng W.Y. (2012) Night-time offshore helicopter operations – identification of contextual factors relevant to pilot performance, Proceedings of the Applied Human Factors and Ergonomics International 2012 Conference. San Francisco, USA, July 21-25 2012.
5. Nascimento, F.A.C., Majumdar, A., Ochieng W.Y. and S. Jarvis (2012) Assessing the hazards of night-time offshore helicopter operations, Transportation Research Board – 91st Annual Meeting. Washington, D.C. USA, Jan 22-26 2012.
6. Nascimento, F.A.C., Majumdar, A., Ochieng W.Y. and S. Jarvis (2011) Safety hazards in night-time offshore helicopter operations, Proceedings of the 37st European Rotorcraft Forum. MAGA Valente, Italy, Sep 13-15, 2011.

Research Project: Fatal Train Accidents and Railway Fatalities in Great Britain

Andrew Evans has produced and circulated an analysis and commentary on fatal train accidents on Britain’s main line railways at the end of every year since 1999. The paper now covers all the principal types of train accidents including collisions between trains and road vehicles.

The end-of-2010 analysis is based on the data from 1967 to 2010. Evans continues to receive positive feedback from readers. As part of this work he keeps his data up-to-date using published data and contacts with Office of Rail Regulation (ORR).

Evans provides input to help maintain the quality of the ORR data.

Publications

1. Evans, A.W. (2011). Fatal accidents at railway level crossings in Great Britain: 1946-2009. Accident Analysis and Prevention, 43(5), 1837-1845.
2. Law, T.H., Noland, R.B. and A.W. Evans (2011). The Sources of the Kuznets Relationship between Road Fatalities and Economic Growth. Journal of Transport Geography, 19, 355-365.
3. Evans, A. W. (2011). Fatal train accidents on Europe’s railways: 1980-2009. Accident Analysis and Prevention, 43(1), 391-401

Research Project: Development of a Data Quality Index for Safety Management Information Systems in Railways

Background

Safety Management Systems (SMS) in railways as well as other industries are fundamental to gather information about safety critical incidents and accidents. They provide the basis for safety analysis and risk modeling that is used for diverse strategic and operational decisions e.g. maintenance planning. SMS collect a wide range of accident and incident datasets from one or numerous railway operators.

These datasets are not necessarily entered by organisations or data analysts with the same understanding of safety and safety management processes. Consequently this affects the quality of data values, its integration and comparability. Likewise data quality issues lead to uncertainties within safety and risk models and weaken the validity.

In order to assess the quality of datasets within a SMS among different accident types and operators, data quality dimensions and indicators need to be defined, measured, analysed and subsequently improved. Herewith, the validity of safety data can be ensured and sources of uncertainties may be adressed. Moreover, safety managers can identify areas of improvement for safety management processes, the safety culture and reporting mechanisms.

Approach / Methodology

This study contemplates the Safety Management Information System (SMIS) operated by the Railway Safety Standards Board (RSSB) in Great Britain. It aims to define data quality dimensions and measures for safety management information and to propose a model of influential factors that are relevant for the railway industry and other industries dealing with SMS. The study provides insights into data quality of empirical datasets on workforce assaults from 2010 to 2012 and compares the subjective perceptions of data quality of 28 transport operators for the same datasets. These subjective perceptions are gathered by a comprehensive interview series among the British transport operators in 2012. These datasets are integrated into a data quality index for the British railway industry using an Analytical Hierarchy Process.

The research thereby gives new insights into the determinants of data quality within safety management systems. It provides a first methodological approach how to measure and compare data quality and identifies areas of improvement. Furthermore, the research project benefits from its conjunctive quantitative and qualitative approach as well as its integration of different research disciplines; such as railways, safety, information and quality management.

Research Project: A Review of Accidents at Road Works in London from 2005 to 2010

Background

Accidents at road work sites have been the focus of multiple studies around the world. However, in the United Kingdom, only a handful of studies have focused on road work accidents on motorways but none have focused on London, the capital city. This report has set out to identify the key characteristics of accidents at road works sites in London, understand where these characteristics differ from accidents at non road works sites and how road work characteristics impact the severity of accidents. Using STATS 19 data, 1138 accidents at road works sites between 2005 and 2010 were analysed. It was found that accidents involve mainly middle aged, male drivers. Also unlike in previous studies, speed was not found to be a significant factor in causing these accidents. These accidents display a number of similarities with accidents not at road work sites including the fact that environmental factors such as weather do not significantly influence the number of accidents.

A number of differences however, were also identified including that most accidents at road work sites occur in Central London as opposed to North London, while accidents at road works have increased over the last five years rather than the falling trend for all other types of accidents.

Using three ordered probit regression models, the impact of accident characteristics on casualty severity was also reviewed, finding that those travelling in cars are more likely to be involved in higher severity accidents, while driver errors such as failing to look properly also significantly increase the odds of being involved in a more severe accidents. Based on these findings and on past literature a number of mitigation measures have been proposed.

Collaboration

This research project was in conjunction with Transport for London.

Research Project: Causal Analysis of Effects of Road Traffic Interventions on Road Safety

Background

In recent years, some researchers have paid close attention to the impact of road traffic interventions on road accidents. Road traffic interventions (RTIs) are policies, enforcements or engineering works that directly or indirectly affect traffic characteristics, such as traffic flow, speed and density, all of which may affect road safety.

The mechanism of how RTIs affect road safety can be very complex depending on the kind of RTIs. For example, as fuel prices and road users’ taxes rise cars will tend to be driven less. It can be hypothesized that since a higher tax leads to fewer miles travelled, the road will be emptier and the traffic speed will be higher. Improved traffic flow condition could in turn affect the occurrence likelihood and severity of traffic casualties. In addition to the complex mechanism, it is also difficult to control for factors confounding the evaluation.

Confounding effects are usually due to individual heterogeneity, time trend, data limitation and other factors not related to the implementation of RTIs. Failure to control for confounders can bias the evaluation result. The superiority of causal models is that it can estimate the effect of RTIs and eliminate confounding effects in one integrated procedure.

Methodology

This study aims to develop a causal method for estimating the impact of RTIs on road safety. First, data for accidents, road characteristics, traffic characteristics and demographic information are collected and organized to aggregated or disaggregated level. Then a causal model, regression or matching method, is chosen and developed by combining with road accident frequency or severity models. Finally, the impact of RTIs is evaluated and suggestions are proposed to improve road safety. Thus far, two case studies have been analysed: the London congestion charge and its impact on accidents; the impact of speed cameras on accidents.

Sponsorship

This PhD is sponsored by the China Scholarship Council (CSC).

Publications

Li, H., Graham, D. and A. Majumdar (2012) The Effects of Congestion Charging on Road Traffic Casualties: A Causal Analysis using Difference-In-Difference Estimation, Accident Analysis & Prevention, 2012, (accepted).

Li, H., Graham D.G. and A. Majumdar (2012) The impacts of speed cameras on road accidents: an application of propensity score matching methods, Proceedings of the 1st European Symposium on Quantitative Methods in Transportation Systems. Lausanne, Switzerland, September 4-7 2012.