Although perhaps not as obvious a route for reducing ongoing costs as physical assets such as rolling stock, track, depots or stops, one of the major ‘everyday’ costs for urban rail undertakings is in managing their electricity demand.
And with ever-growing service hours, more intensive daily operations and greater demands from more complex vehicle systems, the requirement for power on electrically-operated systems around the world is growing. As society as a whole is consuming ever more energy, costs for base supply go up and all of this can easily equal substantially larger bills.
The rail industry has long strived for greater efficiency through traction energy savings. While light rail and metro systems are amongst the most efficient modes of powered transport, where energy use per mile is much lower than other modes, about 80-90% of the energy consumed in real-world operations is for the vehicle’s traction supply. It is therefore in the interest of undertakings and operators to find ways to reduce energy consumption from their day-to-day activities – and one of the best ways of doing this is by lowering the demand from the vehicles.
Innovation through collaboration
A presentation from the University of Birmingham on an innovative modelling programme for light rail energy usage at the 2016 UK Light Rail Conference led to a valuable conversation between the presenters and Trevor Dowens of technical consultancy Ricardo Rail.
Trevor Dowens, Principal Consultant in Ricardo Rail’s Operational Consulting division, explains the background: “After some years of looking for a suitable model that could help inform driver training processes, we met Dr Stuart Hillmansen and his colleague Dr Rob Ellis from the University of Birmingham.”
The University of Birmingham’s Centre for Railway Research and Education (BCRRE) has developed simulation software for energy efficiency, which has been successfully tested on a number of systems in China and also by Edinburgh Trams in the UK. The research was originally developed as part of UK Tram’s Low Impact Light Rail initiative, which informed the modelling process for the programme.
Ricardo Rail and the University began working together after the 2016 Conference, working through various forums to add functionality to the model. Its use for the combined energy modelling process helped inform a suite of driver training and behavioural change processes aimed at making substantial cost savings on the use of traction energy. The result was SmartDrive.
SmartDrive offers driver training informed by the analysis of detailed data of individual routes – and the system as a whole – identifying specific sections (built up as a map over an entire network) and creating optimal trajectories for a tram or LRV in service.
Dowens explains: “In effect it’s a profile of acceleration, coasting and braking – and from that we can more specifically inform the training process, which of course is always the most challenging element due to the changes in driving styles and behaviours required.
“In terms of the optimal trajectory, a typical vehicle’s path is typified by rapid acceleration – this should be a fairly straight line. There’s the longest period of coasting possible, and then steep deceleration at the end – that’s the optimal vehicle trajectory for energy saving.”
He continues: “I can certainly remember during my time at the Tyne and Wear Metro in the early days we had a traction inspector who took it upon himself to place coasting boards around the system. It was an excellent initiative, but it didn’t work very well as there was no science behind it and consequently there was little buy-in from either drivers or senior management.
It was a valiant attempt though, and I’m sure that other rail organisations had the same sort of experiences.” Ultimately, the initiative failed. Through SmartDrive, trials have shown that traction energy costs can be reduced by 10-25%, representing a significant potential saving on an organisation’s energy bill.
Dowens believes that the industry has traditionally favoured technology-led approaches when considering energy efficiency – such as regenerative braking and DAS (driver advisory systems), but this is changing: “Evidence is emerging of a more behavioural and driver-centred approach to economic driving. Both approaches can each save in the region of around 5% of energy costs. But if you combine the two with simple modelling, it can generate significant additional benefits.”
Importantly, the system both empowers and engages the driver, Dowens explains: “There’s no in-cab display, it’s not a DAS and as a result it is compatible with LRT systems that use line-of-sight driving. Not being mandatory, it empowers the driver, by not telling them what to do and when, and depending on the circumstances involved he or she can choose whether to adopt the driving profile required or not, based on performance and other considerations.
“Another benefit is that it gives an improved passenger experience through a smoother and less jerky ride as we’re managing the transitions between acceleration, coasting and braking. Rolling stock maintenance teams will thereby also see improvements in reduced wear and tear through less frequent braking.”
The Ricardo and BCRRE team also claim that the technology is less complex and cheaper than in-cab driver advisory systems – SmartDrive can be implemented far quicker and at a fraction of the price.
From the testing regime, it has been demonstrated how the revised driving profiles adopted are both safer and can enhance operational performance.
The system can also be used to inform point-to-point timings and route design – for example, factoring in the need for smoother driving profiles and regenerative braking it can help the positioning of substations.
“Interestingly, the technology can also reduce energy costs for some Automatic Train Operation (ATO), or driverless, metro systems. Being a naive operator, I always thought that ATO systems and the computers that operated them already had a good measure of energy preservation built in, but in reality there’s still a big margin for improvement. This process can help improve the energy efficiency of existing ATO systems, so there’s another knock-on benefit aside from driver training.”
Creating the model
Dr Stuart Hillmansen from the BCRRE explains some of the testing and subsequent modelling through a trial at Edinburgh Trams, beginning with a basic principle: “Newton’s Second Law of Motion describes everything about rail motion; all simulators have that buried somewhere within them. If you’ve got those calculations right you can solve everything. It’s the most important equation for traction, and it applies equally in Liverpool, in China, on the West Coast Mainline and on all other railways around the world.
“It’s simple stuff, fundamentally, but we’ve spent a lot of time on how you work out the optimal way of driving a rail vehicle between two points – so we can calculate a ‘golden run’.
On the prototyping process, Hillmansen says: “One evening I travelled 80km (50 miles) on an Edinburgh tram overnight. We went back and forth, helping the driver identify the best places to apply the power and the brakes.
“We came up with a simple way of informing the driver of what to do and when, including target speeds. We have a simple driver advisory system that we can use in training, but we’ve already got a very good computer in the cab, the driver’s head.
“The Edinburgh Trams system isn’t as big as most urban networks, so they’re already intimately familiar with the routes. We just make use of that computing power rather than adding additional equipment into the cab.”
Two test runs of the full line between Edinburgh Airport and the city centre were completed, one with the existing timetable and one with the optimised version to compare savings in various sections. “With optimal running you’re not hitting such high speeds and braking a little later – you use more acceleration, a lot of coasting, and brake quite heavily in the latter part of the journey,” comments Hillmansen. “All this was known 100 years ago, so we’re not reinventing the wheel, but equally some of these driving techniques have been forgotten and I think they’re worth revisiting with the science behind them to explain how they work.”
When the energy usage was added up from the test runs, savings of up to 10-20% were found from the optimal runs. Stuart adds: “The drivers ‘bought in’ to the concept, and tended to engage more once they believed it would work. We also saw slightly lower journey times overall, so it’s very good from that point of view as well.”
A second field test looked at the high-speed timetable, as Edinburgh Trams was very conscious of trying to reduce its end-to-end journey times. The drivers were already going faster in the central section, and Stuart had less to work with in terms of contingency time: “This is an example where the standard driver would coast a little bit, and then re-motor and then another coast and then brake, and then a little bit of coast and brake again, so we can get rid of this area.
“The interesting thing about the new model is that the driver starts coasting around 800m out of the station. If you look 800m along the track it’s a very long way, and we’re telling the driver to stop using the power at this point and go into coasting. The drivers initially looked unconvinced, but the trams are quite aerodynamic, they’re quite heavy, they’ve got low rolling resistance and they will go a long way without any power. That’s part of the driver education: convincing drivers that this is how the vehicle can be used differently.”
The second trial saw up to 13% savings in traction energy on the higher-frequency timetable. Dowens adds: “Having managed operational front line staff for most of my working life, the challenge is the management of change. Behavioural change, and the training and development of drivers, is crucial to this. Once the fix has been put in, sustaining the fix brings its own challenges.
“Also, if you’ve got regenerative braking, the profiles are different in terms of putting the energy back into the overhead lines. When we do the optimisation model, we look at the total energy of the system and you will have a slightly different answer with regenerative braking.
“Edinburgh’s trams don’t regenerate back into the network, you would get much greater savings if they did and you’d come up with slightly different braking trajectories. Ideally, if you’ve got regenerative braking you want to do most of your braking just using the regenerative system and not friction braking.”
Achieving effective change involves proactive engagement with drivers, trainers and management, Dowens says, and to some extent this depends on the organisational culture in terms of how receptive staff are to these initiatives. Without positive engagement at all levels, he argues, there’s little chance of moving forward collectively and positively.
He describes the process: “The first stage is data gathering. That draws upon vehicle and infrastructure data to inform the model and come up with a set of theoretical trajectories.
“After the initial modelling is complete, a testing programme is designed on a route-by-route basis to ensure the trajectories work in the operational environment. If they don’t, the model is adjusted. At this point we’re also looking more generally at safety and performance: Are the new driving profiles safe? Are the braking curves achievable? Does operational performance suffer?”
Practice runs inform the location of lineside signage in relation to the rest of the urban environment, as the drivers require some visual clues for when to coast, for example. Simple signage can be placed on overhead line structures and so forth, but its exact placement is crucial in terms of both visibility and reducing vulnerability to outside interference.
The next step is training course development, created on a bespoke basis and taking into account existing professional driving policies and specific rules and regulations. Theoretical classroom training is then undertaken, explaining and understanding the science behind the theory, followed by practical on-route training.
Some organisations may choose to have a steering group involved in managing the process and monitoring its implementation to ensure the envisaged benefits can be realised. Metrics can empirically demonstrate anticipated energy savings, driver empowerment, and customer experience and this may inform infrastructure, rolling stock or maintenance changes that require refinement. The ideal end results give improvements across the whole business: one possible counter-intuitive outcome is that driving styles with more coasting often give shorter point-to-point times than typical ‘continuous power’ styles.
Dowens concludes: “There must be a commitment and support from senior management to create the right environment for behavioural change, otherwise these initiatives will just wither on the vine. Keeping managers involved in the process will encourage continuous improvement and driver feedback and support in continually refining the model is key. Only then will you see the dramatic improvements in overall driver performance and energy usage.
“We want to make drivers more professional, we want to encourage pride in terms of what they’re doing and we want to make them more focused on what they’re doing. This process can do all that while also delivering efficiencies across the organisation.”
Feature originally published in November 2017 TAUT (959).