Research: COMET Preventing Human Failures in Train Operations and the OCC

There are multiple reasons why metros are particularly affected by human failures, including operations and/or maintenance occurring 24-hours a day, human judgement involved in multiple safety critical situations, and the potential for unpredictable overtime. Across COMET metros, human failure caused the majority of collisions and derailments. Miscommunications and multi-tasking are considered the top two most influential factors on human failures.

The study focused on active human failures that directly lead to incidents. Active failures are categorised into errors and violations, which can be further sub-categorised by the underlying motive, intention, and/or frequency. Understanding active failure types supports greater understanding of the root causes of failure.

The study reviewed strategies for preventing human failure including process, environment, and people-based solutions. One example for preventing human error is assessing fitness for duty among train drivers. Half of metros reported use of a checklist or declaration, with some metros use technology to address fatigue including monitoring attentiveness and encouraging sufficient rest for employees before their shift. While many of these measures are focused on train drivers, increasing automation within metros means that such measures are becoming just as important for OCC staff. Opportunity areas for metros to reduce operational risk of human failures are discussed, supported by good practices and future initiatives conducted by COMET metros.

Research: Driver Training

Training drivers and maintaining their skills and knowledge are significant efforts for almost all metros. Metros need to implement adequate selective recruitment processes to find suitable candidates to become drivers.

The core of initial driver training programmes amongst metros is largely similar, averaging at 100 working days. Programme length depends on a number of factors, including external vs internal recruitment, training facilities, and metros’ expectation towards the role of drivers. The content of training courses used by metros was explored by the study as well.

Duration of Initial Driver Training by Teaching Structure

Apart from the driver initial training, the study also reviewed the frequency of the recurring driver training. By comparing the duration of the driver training to the reliability performance, a correlation was identified between longer training and fewer staff-related incidents causing delays.

Training methods are evolving, as technological advances allow for a greater reliance on simulators to enable drivers to gain experience and confidence in a controlled environment. The developing dependence on mobile devices was identified as an opportunity to integrate more mobile technology into recruitment and training.

Research: Multifunctional Staff

Nova members have identified a need to innovate to increase staff productivity levels, and asked RTSC to investigate how metros around the world have used multifunctional staff. A wide variety of multifunctional roles were identified, classified into six broad types as shown below.

multifuncitonal roles

The best multifunctional staff roles fill in what would otherwise be unproductive time, with productive activity. This is often accomplished by matching functions that need to be done at separate times of day or functions that can be slotted in between other activities in a single location, such as light maintenance within stations.

Multifunctional working also has an important role at increasing staff satisfaction. By combining tasks, staff have the opportunity to work in a more varied and interesting role. This can improve the attractiveness of the metro as an employer and improve staff motivation. For example, one metro recorded reduced absenteeism among their most multifunctional staff. Multifunctional roles can also create a career progression – especially for staff who are technically excellent but do not necessarily want to manage other people.

Research: Unattended Train Operation

RTSC’s recent research into unattended train operations (UTO) has investigated the role of human operational support on UTO lines. In doing so, a key finding relates to the use of attendants or train captains – metro staff members who are based in the passenger car, rather than in a separate driver’s cab. Formal definitions of Grades of Automation as in IEC-62290 assume that if a staff member is onboard, they are fulfilling a necessary role, in operating the train, as in London’s Docklands Light Railway where attendants close the doors. The assumption is therefore that lines capable of being operated unattended (Grade of Automation 4) are operated unattended. This research study has found that in fact some metros with GoA4 lines actually use attendants on all trains for other reasons, for example to provide customer service. This has led RTSC to describe this type of line as ‘Attended GoA4,’ reflecting the fact that it fits the specification for a GoA4 line as described in IEC-62290, but is not being operated unattended. The diagram below illustrates how ‘Attended GoA4’ automation fits in with the grades of automation.

Adapted from IEC-62290 for the purposes of demonstrating differences in the real-world operational application of the formal Grades of Automation (GoA1-4) defined in the standard.
Adapted from IEC-62290 for the purposes of demonstrating differences in the real-world operational application of the formal Grades of Automation (GoA1-4) defined in the standard.

The case study explored in more detail the actual staffing levels used or planned on participating metros’ automatic lines. The total number of operational staff was compared with the number of assets (stations and trains) in service, providing metros with a useful benchmarking metric that normalised for differences in line length and service level. A typology of staffing models was also developed and linked with associated staffing levels. This work on staffing was complemented by investigation into the technologies required to enable automation; their costs; and metros experiences with their reliability.

This research was presented as a poster at the 94th Transportation Research Board of the National Academies Annual Meeting in Washington DC, in 2015.

Research: Train Driver Productivity

A Nova research study on train driver productivity aimed to identify the most important factors that influence driver efficiency, understand what methods operators have successfully used to modernise restrictions to working arrangements and identify the scope for metros to modify the most important constraints, rules and parameters that have a negative effect on both driver productivity and costs.

The study found that metros need to have sufficiently flexible labour rules to achieve higher levels of driver productivity. Correlating the level of working time flexibility with driver productivity showed that less restricted metros were more productive than metros that face stricter constraints. Organisations with part time drivers and / or the ability to utilise split shifts were associated with higher levels of driver productivity and increased effectiveness of driver time at work (better able to cover ‘peaks and troughs’ in service). Moreover, it was found that variable shift lengths were arguably more effective than split shifts for metros with a flatter service profile, allowing metro managers to adjust shift schedules as necessary and to avoid the build up (unproductive time) of staff during less busy and off peak periods. In the long term, savings from increased flexibility could more than offset the higher driver wages associated with greater flexibility. Increasing automation of train services through automatic train turnaround, remote booking-on for drivers and driverless trains can positively influence driver productivity and allow driver roles to be deployed more effectively through more customer facing roles.

It is hoped that the information contained within this study can be used to support negotiations with labour unions, government and other key stakeholders by showing evidence of how “imposed” restrictions may impact an organisation’s ability to improve and manage productivity.