Rail Safety And Standards Board Research Articles

Developing a novel composite index for monitoring occupational health and wellbeing: A case study in the rail sector in Great Britain

The high economic and social burden of poor physical and mental health among workers calls for strategies to support and protect workforce health. Monitoring occupational health and wellbeing through the consistent and systematic use of data is considered to contribute substantially to designing evidence-based policies that improve outcomes for workers. The purpose of this paper is to expand on current efforts to enhance monitoring practices by introducing a novel approach to developing a composite Health and Wellbeing Index (HWI). HWI can combine a range of areas (for example, sickness absence, exposure to health risks, key health and mental wellbeing) based on valuation methodologies from health economics. The cost of different outcomes to companies, individuals and society is identified using Willingness to Pay (WTP) and Quality Adjusted Life Year (QALY) methods. These costs are compared with estimates of the value society attaches to preventing a fatality to arrive at weightings that reflect each HWI component’s monetary equivalence to the cost of a fatality. Based on this method, different components are transformed into comparable units and are added together to calculate a headline metric. A version of HWI that is tailored to the needs of the rail sector in Great Britain was developed in collaboration with the Rail Safety and Standards Board (RSSB) to drive improvements in the sector’s performance. The new index can be used to track progress over time, benchmark against best practice, identify areas for improvement and develop a robust evidence base for policy design and investment decisions.

Blockchain-hosted Data Access Agreements for Remote Condition Monitoring in Rail

Advances in sensor technologies, remote authentication, and high bandwidth data networks mean that Remote Condition Monitoring (RCM) systems are now an essential 'Internet of Things' (IoT) resource for the efficient operation of railway infrastructure. However, the full potential of the big data generated by these systems has yet to be realized. RCM data within the industry is typically collected and used in silos, with limited possibility of exploitation across system boundaries. In 2013, the Rail Safety and Standards Board (RSSB), on behalf of the GB Rail industry, established a cross-industry research program, T1010, which aimed to build stronger cooperation between stakeholders and enable sharing of RCM data. Building on the outputs of T1010, this work explores the use of blockchains and smart contracts in the automation, in an auditable and tamper-proof way, of commercial agreements for RCM data transfers in rail. By removing the limitations of paper-based agreements, we aim to enable innovation in shared business processes and stimulate the market for RCM data in rail. Leveraging existing smart contract-based schemes for trading and sharing IoT data over blockchain networks, we identify suitable methods for the enforcement of agreements, and ensure fair cost attribution between stakeholders, without a Trusted Third Party. The outline of a blockchain-based RCM data audit framework is presented, appropriate data access agreements and accounting models are specified in detail, and three permissioned blockchain platforms (Hyperledger Fabric, Sawtooth, and Iroha) have been analysed for their suitability for implementation. Finally, the paper outlines planned future work around validation of the tools based on two industrial use cases: monitoring systems for unattended overhead line equipment and axle bearings.

Object Detection at Level Crossing Using Deep Learning.

Multiple projects within the rail industry across different regions have been initiated to address the issue of over-population. These expansion plans and upgrade of technologies increases the number of intersections, junctions, and level crossings. A level crossing is where a railway line is crossed by a road or right of way on the level without the use of a tunnel or bridge. Level crossings still pose a significant risk to the public, which often leads to serious accidents between rail, road, and footpath users and the risk is dependent on their unpredictable behavior. For Great Britain, there were three fatalities and 385 near misses at level crossings in 2015–2016. Furthermore, in its annual safety report, the Rail Safety and Standards Board (RSSB) highlighted the risk of incidents at level crossings during 2016/17 with a further six fatalities at level crossings including four pedestrians and two road vehicles. The relevant authorities have suggested an upgrade of the existing sensing system and the integration of new novel technology at level crossings. The present work addresses this key issue and discusses the current sensing systems along with the relevant algorithms used for post-processing the information. The given information is adequate for a manual operator to make a decision or start an automated operational cycle. Traditional sensors have certain limitations and are often installed as a “single sensor”. The single sensor does not provide sufficient information; hence another sensor is required. The algorithms integrated with these sensing systems rely on the traditional approach, where background pixels are compared with new pixels. Such an approach is not effective in a dynamic and complex environment. The proposed model integrates deep learning technology with the current Vision system (e.g., CCTV to detect and localize an object at a level crossing). The proposed sensing system should be able to detect and localize particular objects (e.g., pedestrians, bicycles, and vehicles at level crossing areas.) The radar system is also discussed for a “two out of two” logic interlocking system in case of fail-mechanism. Different techniques to train a deep learning model are discussed along with their respective results. The model achieved an accuracy of about 88% from the MobileNet model for classification and a loss metric of 0.092 for object detection. Some related future work is also discussed.

Learning From Accidents: Machine Learning for Safety at Railway Stations

In railway systems, station safety is a critical aspect of the overall structure, and yet, accidents at stations still occur. It is time to learn from these errors and improve conventional methods by utilizing the latest technology, such as machine learning (ML), to analyse accidents and enhance safety systems. ML has been employed in many fields, including engineering systems, and it interacts with us throughout our daily lives. Thus, we must consider the available technology in general and ML in particular in the context of safety in the railway industry. This paper explores the employment of the decision tree (DT) method in safety classification and the analysis of accidents at railway stations to predict the traits of passengers affected by accidents. The critical contribution of this study is the presentation of ML and an explanation of how this technique is applied for ensuring safety, utilizing automated processes, and gaining benefits from this powerful technology. To apply and explore this method, a case study has been selected that focuses on the fatalities caused by accidents at railway stations. An analysis of some of these fatal accidents as reported by the Rail Safety and Standards Board (RSSB) is performed and presented in this paper to provide a broader summary of the application of supervised ML for improving safety at railway stations. Finally, this research shows the vast potential of the innovative application of ML in safety analysis for the railway industry.

Wheel-rail creep force model for predicting water induced low adhesion phenomena

A computationally efficient engineering model to predict adhesion in rolling contact in the presence of water is presented which may be implemented in multibody dynamics software or in braking models to study train performance and braking strategies. This model has been developed in a project funded by the Rail Safety and Standards Board (RSSB) and Network Rail. It is referred to as the water-induced low adhesion creep force (WILAC) model. The model covers a wide range of conditions from dry over damp to wet. Special emphasis is put on little amounts of water which can cause low adhesion without any oil or grease. Such conditions may be encountered in humid weather or at the onset of rain. The model is parameterised based on experimental results from a tram wheel test rig. Adhesion values as low as 0.06 are observed at high creep with only wear debris and little water present in the contact. The model results also agree with experimental data from locomotive tests in dry and wet conditions.

Efficiency Benefits from the Deployment of a Novel Flywheel Solution for Non-Electrified Lines

This paper describes the authors’ investigation into deploying durable, low cost electric flywheel technology on rail vehicles by means of simulation. Diesel engines currently power trains on non-electrified lines, over half the UK rail network, so the Rail Safety and Standards Board (RSSB) commissioned research to devise efficiency improvements. Substantial benefits in the form of diesel traction energy saving and braking energy recovery were calculated, amounting to 43% for regional passenger services. Further, the flywheel technology is traction-type independent: it could be key to unlocking discontinuous electrification or to improve the effectiveness of any prime mover. A follow-on hardware demonstration project has just commenced.

Railway suicide in England and Wales 2000-2013: a time-trends analysis.

BackgroundIn 2010, the “Tackling Suicide on the Railways” programme was launched as a joint initiative among Network Rail, the Samaritans and other key organisations such as the British Transport Police and train operators to achieve a 20 % reduction in railway suicides from 2010 to 2015 in Great Britain. We report the most recent age and sex specific trends in railway suicide in England and Wales from 2000 to 2013 and examine whether the initiative’s target reduction in railway suicides is likely to be achieved.MethodsPopulation data and suicide mortality data (all methods combined and railway) for England and Wales were obtained from the Office for National Statistics (ONS) and used to calculate age and gender specific rates for deaths registered from 2000 to 2013. Data on railway suicides were also obtained from the Rail Safety and Standards Board (RSSB) and compared with ONS data. We used joinpoint regression to identify changes in suicide trends across the study period.ResultsThe railway was used in 4.1 % of all suicides in England and Wales (RSSB data were similar to ONS data for most years). Suicides in all persons from all causes decreased from 2000 to 2007, with small increases from 2008 until 2013; this rise was entirely due to an increase in male suicides. Railway suicide rates increased over the entire study period; the proportion of railway suicides in all persons increased from 3.5 to 4.9 % during the study period. This trend was also mainly driven by increases in male suicides as female railway suicide rates remained steady over time. The highest age specific railway suicide rates were observed in middle aged men and women. Although there was no conclusive evidence of an increase in ONS railway suicides, RSSB data showed a statistically significant increase in railway suicides in males from 2009 onwards.ConclusionThe continued rise in male railway suicide in England and Wales is concerning, particularly due to the high economic costs and psychological trauma associated with these deaths. The initiative’s target of a 20 % reduction in railway suicide is unlikely to be achieved.

Localising Risk Estimates from the RSSB SRM

The Rail Safety and Standards Board (RSSB) Safety Risk Model (SRM) provides a structured representation of the causes and consequences of potential accidents arising from railway operations and maintenance on the UK railway network. It consists of a series of fault tree and event tree models representing 120 hazardous events that collectively define an overall level of risk on this system.Recently there have been several RSSB led projects looking at how to take the national results from these models and apply them at a local level. This involves breaking the network up into smaller regions and using the localised data to estimate the risk. There are two main ways of doing this: either scaling the national profile using appropriate normalisation data or by taking the calculations used at the national level and applying them to the localised data.The paper will look at the tools RSSB has developed over the years to carry out these assessments. The scaling of the SRM figures from a national profile to a localised profile is done via the SRM Risk Profile Tool. The application of the SRM calculations to localised data to calculate a localised profile can be done via a tool called the Network Modelling Framework (NMF) Safety Module. The purpose of this paper is to discuss how these tools work, their limitations, some of the challenges that have been encountered and current research aimed at developing a scalable ‘GeoSRM’.

Assessing the risk to railway structures from train impact

Designers of railway structures frequently find that they have difficulty meeting the requirements for design of structures to resist impact loads from a derailed train. Potentially such loads can be extremely high, especially where a very substantial proportion of the train mass can be mobilised at line speed, as was the case for the Eschede incident in Germany. It is rarely economical to design structures for such extreme loads and design standards such as BS EN 1991-1-7 set impact load values that are intended to provide a minimum level of robustness for safety and economy. However, the safety benefits compared to the cost of providing sufficient resistance, even for the loads in design standards, can be difficult to justify in some cases. Risk assessment can be used in such circumstances to demonstrate achievement of an acceptable level of safety at reasonable cost. Application of the risk assessment methodology in UIC leaflet UIC-777-2R is described, highlighting the availability of Rail Safety and Standards Board (RSSB) derailment risk data and RSSB guidance on taking safe decisions.

Development and validation of a wheel wear and rolling contact fatigue damage model

This paper summaries the development of a damage model to predict the deterioration rates of the wheel tread in terms of wear and rolling contact fatigue (RCF) damage. The model uses a description of a fleet's route diagram to characterise the duty cycle of the vehicle in terms of curve radius, cant deficiency and traction/braking performance. Using this duty cycle a large number of vehicle dynamics simulations are automatically conducted to calculate wheel–rail contact forces and predict the formation of wear and RCF damage, using a combination of the Archard and frictional energy-based (Tγ) damage models.The damage models have been validated using observation data (wear rates and maximum observed RCF damage) acquired from a range of vehicle fleets in Great Britain (GB). Results from the validation of the model are presented along with a review of the wheel turning and observation data.A piece-wise linear regression is fitted to the wear and RCF parameters predicted from the model to determine the damage rates for each wheelset type on the vehicle. These damage rates are used within the recently developed Wheelset Management Model (WMM) to describe how the attributes of the wheel (i.e. wheel diameter, profile shape and tread damage) deteriorate over time and trigger a maintenance or renewal activity when the condition of the wheel matches a particular limiting value.This work formed part of the rail industry research programme managed by the Rail Safety and Standards Board (RSSB), and funded by the Department for Transport, to increase the rolling stock functionality of the Vehicle Track Interaction Strategic Model (VTISM) tool.

Modelling Uncertainty in the RSSB Safety Risk Model

The Rail Safety and Standards Board (RSSB) Safety Risk Model (SRM) provides a structured representation of the causes and consequences of potential accidents arising from railway operations and maintenance on the Great Britain railway network (GBRN). It consists of a series of fault tree and event tree models representing 120 hazardous events that collectively define an overall level of risk on this system.In 2003–2009, RSSB jointly undertook research with Strathclyde University into developing a method of calculating uncertainty that could be used on the SRM's models. Since these models are built using fault and event tree logic, they consist of many branches of related variables. This creates a significant challenge in terms of calculating uncertainty, as there are many interrelated factors to take into account when constructing the uncertainty model.The research has looked at various ways of eliciting correlation and covariance matrices that can be used to simplify the uncertainty analysis while still taking account of the interdependency between variables. These methods are used in conjunction with a Monte-Carlo simulation to sample from distributions and produce data samples that can be used to populate the SRM models. These models are then run many times using the samples, in order to build up distributions of expected risk results. They can then be used to determine confidence intervals on the risk figures.This research is now complete and a case study on one of the hazardous events of the SRM has been carried out. The purpose of this paper is to discuss the application of the uncertainty models to the SRM and to present the results that have been obtained using this uncertainty method.

Developing safety signs for children on board trains

Every year a number of young children are injured as a result of accidents that occur on board trains in Great Britain. These accidents range from being caught in internal doors, through to injuries caused by using seats. We describe our efforts to design a new set of safety signs in order to help prevent the occurrence of these types of accident. The research was funded under a Rail Safety and Standards Board (RSSB) managed UK Department for Transport research programme and was carried out in collaboration with Loughborough University. The study involved analysis of industry accident incidence data and running a set of classroom discussions with young school children (aged 5–10, n=210). The classroom discussions initially involved showing them examples of a new design prototype sign alongside existing train signs and gathering the requirements for new designs. A second set of classroom discussions with these children was used to evaluate the new signs based on the outcomes from earlier discussions. We describe our findings alongside a set of outline guidelines for the design of safety signs for young children. A final section considers the main methodological and other lessons learnt from the study, alongside study limitations and possibilities for future research.