Research in progress
Small steps towards ‘super-fast’ freight | Pursuing a step change in monitoring safety-critical communications | A smarter approach to managing traction energy
Work on a specialised rig will explore the aerodynamic risk from faster freight trains and how to mitigate it.
For many decades, all freight traffic on the GB network has been limited to a maximum speed of 75 mph. Even as a new generation of freight locomotives and wagons is entering service with the capability to haul freight at higher speeds, this limit remains in place.
While there are various barriers to faster freight, one reason for the blanket restriction to 75 mph lies in aerodynamic risk. A freight train travelling at higher speeds creates stronger slipstream and pressure waves, potentially increasing risk on the network—particularly when going through stations.
Following an initial review (project T1303), RSSB has identified areas for further investigation, where developing the evidence base may unlock higher speeds. Aerodynamic forces at freight speeds above 75 mph have not been assessed, as there was no previous need. There is, therefore, no evidence to say whether aerodynamic forces created by travelling faster than 75 mph are acceptable or not.
Using the University of Birmingham’s Transient Aerodynamic Investigation rig, which is a purpose-built testing facility for examining the transient aerodynamics of moving vehicles, this work will assess how the resulting aerodynamic forces compare with the allowable levels and determine if and how freight train speeds above 75 mph can be accommodated.
Stay up to date on the progress of the project at rssb.co.uk/research-catalogue (search for COF-ARF).
To discuss the project, contact Aaron Barrett, Lead Research Analyst:
Aaron.Barrett@rssb.co.uk
Using AI could lead to more thorough, objective assessments and targeted interventions.
Substandard safety-critical communications (SCC) can be a factor in accidents involving trains or rail workers. Rail companies are required to monitor frontline staff’s adherence to SCC criteria to ensure the safe running of the railway.
However, SCC data is currently processed manually, and since monitoring is labour intensive, only a small proportion of all communications can be listened to and assessed. This ultimately reduces the fairness and soundness of these assessments as well as the opportunity to detect and act on existing issues and emerging trends.
This research looked at the potential for artificial intelligence (AI) to monitor and analyse a far greater volume of spoken SCC on a routine basis.
The project developed a proof-of-concept (PoC) system to assess the technical and economic feasibility of using AI to monitor the quality of SCC. The PoC uses two types of AI. First, it automatically transcribes SCC audio into unstructured text using automatic speech recognition (ASR) technology. Second, it processes the transcribed text using natural language processing (NLP) to assess adherence to SCC criteria. A simple dashboard was also developed to showcase how users could interact with the system’s outputs.
The PoC demonstrated that this combination of fine-tuned ASR and NLP technologies could successfully monitor SCC in real-world situations. The system successfully picked up adherence to a selection of safety-critical criteria, albeit not all criteria covered by SCC guidance protocols.
The technology is not yet mature enough to completely replace human reviewers, but by monitoring SCC at a large scale, it could identify and prioritise training and development needs. The results also showed that AI could significantly reduce subjective bias by standardising SCC monitoring processes.
The overall goal was to access the feasibility of using AI to monitor [SCC]. The project demonstrated that there are widespread benefits in using AI for this purpose and that continued development would yield positive results. There is a strong indication that this will have a huge impact on the number of communications and a large reduction in unconscious bias.
A second phase of the research, co-funded by Network Rail, is currently underway to improve and further test the PoC. It will also identify the steps needed to take the system from a PoC to a trial-level version.
Stay up to date on the progress of the project at rssb.co.uk/research-catalogue (search for T1261).
To discuss the research, contact Liz Davies, Professional Lead, Data and Modelling Research:
Liz.Davies@rssb.co.uk
Can we develop a real-time energy management system to better manage supply and demand of power on electrified routes?
Parts of the existing electric traction power network are already running at full theoretical capacity during times of peak demand. Demand for electrical power is likely to increase as more electric, battery, and multi-mode trains join the network, so the challenge will become ever more pressing.
Predicting and managing this capacity more accurately and closer to real time would lead to a smarter use of the power available in the rail grid and, in the future, the power available in batteries on trains.
This research will explore the feasibility of developing an intelligent energy management system for the Network Rail Western Route. It will also produce a cost-benefit analysis and a plan for trialling and deploying such a system.
The Western Route has been proposed because its Network Rail electrification equipment has advanced data communication systems, which should provide real-time power supply capability data. Power supply information could feed into an energy management module to be added to existing traffic management systems, which could then link to the Connected Driver Advisory System.
The project will also consider the feasibility and benefits of adapting the proposed system so that it can be rolled out on other electrified routes.
Stay up to date on the progress of the project at rssb.co.uk/research-catalogue (search for T1270).
To discuss the project, contact Mark Hanham, Research Analyst:
Mark.Hanham@rssb.co.uk
