After Galicia, what now for rail safety? What the experts say…
As with all rail disasters, the horrific derailment in Galicia, northern Spain last month in which 80 people perished has triggered a whole host of questions about how to stop it happening again. Why was the driver going at more than double the speed he should have been? How far can engineering go towards eliminating human error? Will politicians be prioritising investment in rail safety technology now? How might this latest tragedy pose new challenges for rail engineers? Such complex questions require well-informed answers so we turned to the experts for their views.
Senior members of the Institution of Rail Signal Engineers (IRSE), the Institution of Engineering and Technology (IET), the Institution of Mechanical Engineers (IMechE), and leading developer of train safety technology, Siemens, have all kindly shared their knowledge and opinions with us.
But first, for those of us who aren’t specialists in rail engineering, it’s useful to get a handle on some of the terminology. Clive Kessell, Past President and Hon. Fellow of the IRSE, explains that the European Commission is already in the process of rolling out a major initiative called ERTMS (European Rail Traffic Management System). Says Kessell, “ERTMS has two component parts – ETCS (European Train Control System) and GSM-R (Global System for Mobiles – Railways) which is an adaptation of the public GSM 2G system. The GSM-R is part of the communication system that relays information to the driver via the onboard ETCS kit from the infrastructure ETCS kit on the tracks. When both cabin and tracks are fitted with ETCS, safety levels are extremely high because if a driver fails to respond to a signal to slow down, the train will brake automatically and the driver cannot override this.”
The train involved in the Galicia derailment was in fact carrying ETCS kit in the cabin, but had left the High Speed section of track which was fitted with ETCS equipment, and entered a transition area onto conventional tracks which were only equipped with the less sophisticated ASFA (Aviso de Seňales y Frenado Automático) system.
Professor Roger Kemp of the IET explains, “ASFA is at least 20 years old and has two functions. If a driver is about to go through a red signal, the system will brake the train automatically. Its second function is advisory. A signal is communicated and the driver has three seconds to push a button acknowledging it, but he can then go on to ignore it. There is no automatic driver override. However, according to a report in Spanish national newspaper, El Pais, since the accident the authorities have modified the ASFA system on the fatal curve outside Santiago de Compostela so that it will now also automatically cause an over-speeding train to brake.”
Modernisation takes time
So how quickly will older, less automated systems like ASFA be replaced with ERTMS? Philippa Oldham, Head of Transport at the IMechE answers, “The accident in Galicia will provide greater impetus to ensure that the planned roll-out of ERTMS can be achieved. ETCS would have prevented the derailment.” However, she cautions that for ERTMS to be implemented and functioning across the whole of Europe will still take a number of decades. Kessell agrees, commenting, “ERTMS is mandated for all new High Speed lines in Europe. Designated regional main lines – known as TENs or Trans European Networks – are also mandated for ERTMS, but as these already exist they will require a refit, which will take many years.”
The initiative is intended to harmonise safety systems to facilitate cross-border traffic without having to fit rolling stock with multiple systems from different countries. The buzz word for this harmonisation is ‘interoperability’ and according to Kessell, achieving it is one of the main challenges for the developers and manufacturers of ETCS systems.
“Each developer, each rail operator in each country wants its own minor variations and getting them to standardise has posed horrendous problems and left development lagging behind the GSM-R element. The ERTMS Users Group – part of the European Commission – has been battling this ‘bespokery’ for much of its time and the only real solution is for railways to adopt a near common set of operating rules and associated signalling principles. All of this takes time and to date only the Danes, as far as I am aware, have bought into this fully.”
A spokesman from Siemens agrees that standardisation of equipment as a hurdle, commenting that “the sector is intensively working to further improve the consistency of the ETCS/ERTMS specification, identifying gaps and reducing the room for interpretation. A main challenge for engineering firms is related to the safety approval process, especially for the ERTMS equipment onboard trains travelling in several different countries. Here we realise that rail operation along trans-national corridors and the national responsibility of the authorities require further harmonisation of rules and processes”.
Will Galicia act as a catalyst?
But surely an incident that saw so many dead and many more injured and maimed will speed things up? Oldham predicts, “There will be political pressure to speed up the ERTMS programme but there are limits – such as the availability of engineering design resources as well as the time required for the installation itself, both trackside and on trains. The major impact will be in Spain and with the Spanish economy short of funds for public projects, any real impact will be limited. Those areas where the system is installed but only partially commissioned may be speeded up. ERTMS is not a cheap system to install and any change to the programme already planned is likely to significantly increase the cost.”
We asked Oldham if she thought the engineering firms producing ERTMS systems would now be in for a busy time, to which she responds, “This will be a busy time for them, but initially this will be requests for quotations and information about the implications of installing ERTMS. Their business level is determined by the general level of investment in major renewals and new lines. This investment will continue but will be greatest where there is a significant private sector investment, such as in the UK.”
It’s worth noting that although ERTMS is a European initiative, the technology is used in other countries as well, for example Taiwan and Australia. Although Japan already has its own excellent automatic train control system, it has expressed some interest in using ERTMS in the future. Siemens supplied the first ever ERTMS projects in China and Saudi-Arabia and has references in India, Algeria, Morocco, Tunisia, Turkey, Australia and New Zealand.
Although ERTMS can and will be installed on conventional track, until the roll-out across Europe is complete, individual countries will continue to use their own existing train control and signalling systems. For example, the UK uses TPWS which is installed at potentially hazardous sites such as the notorious accident black spot on the 50mph curve at Morpeth near Newcastle, France uses the highly efficient TVM430 system (which incidentally covers most of the British section of the Eurostar High Speed rail track), and Germany uses a system named LZB.
The question of human error
As safety systems continue to evolve, we asked, “How far can engineering go to eliminate human error?”. Kessell replies, “If everything is working OK, technology will be at least 99.8% foolproof. The risk comes when something fails and you work in what is called ‘degraded’ mode. If something fails, the system will stop the train and then you’ve got to have some procedure, some process that enables that train to move without the supervision of the failed system. In theory, from a control system point of view, rail travel is going to get safer and safer. But, that doesn’t stop rails from breaking or axles from splitting in two. A railway is more than just a control system.”
Oldham’s views back this up. “Improved train protection signalling systems are important in mitigating the catastrophic risk of train collision, but statistically this is a lower cause of death and injury than level crossings, infrastructure failures (maintenance related), structures and earthworks problems (weather and maintenance related), mechanical failure of trains and the platform/train interface. Most passenger injuries on railways are slips, trips and falls. Most deaths on or about the railway are suicides, and the growth of technological solutions to help spot and manage these is on-going, but not, of course, the complete answer. However, I do see there being a rise in the amount of automation used on trains to reduce driver error and to assist with remote condition monitoring and automated inspection techniques.”
Professor Kemp’s opinion on technology eliminating human error is, “I don’t think it [engineering] can completely. It’s relatively easy to produce a system that will control and monitor the speed of a train when it’s going along normally, but there will always be some situations when it takes a different set of abilities to control the train.”
In the driver’s seat
He goes on to explore another avenue of thought which centres around the sometimes impossibly high expectations placed on drivers and the issue of disorientation. “The challenge isn’t human error as such, it’s designing a system that doesn’t put humans in an impossible position.”
He then directed us to his blog, written for Lancaster University in his other role as Professorial Fellow of Engineering there. In relation to the derailment in Galicia, he gives this fascinating insight about what might have gone wrong. “In this accident we have to ask whether the driver really knew where he was on the route (or, in more academic language, was he suffering from a deficit in situational awareness).
“The new line passes through difficult terrain and has 31 tunnels, totalling 30 km, and 38 viaducts. The last tunnel before Santiago ends less than a kilometre before the new line joins the old and goes into the 500m radius curve. A week after the accident it emerged that the central control room had phoned the driver as he was approaching Compostela. With the distraction of an administrative phone call, a changeover of the signalling system and a line forever running into and out of tunnels, would it be surprising if the driver had a situational awareness deficit?”
Kessell puts forward a very similar suggestion as to the possible cause of human error at Galicia, positing, “He [the driver] may have lost orientation and wasn’t quite sure where he was. When you ride in the front of a cab on these High Speed lines it does get a little bit monotonous because they are reasonably straight, and in Spain they tend to go right across contours. Tunnel, viaduct, tunnel, viaduct is a fairly typical scenario. He may have become disorientated and forgotten where he was.”
Another tragedy which illustrates some of the difficulties facing drivers and the ‘unfriendliness’ of some of the systems they have to deal with is the Ladbroke Grove incident in the UK back in 1999. The driver, along with 30 others, died when he failed to stop at a red signal and his commuter service collided with an express train. The Cullen Report into the incident found that the driver, Michael Hodder – a 31 year old father-of-two – was only partly to blame. The report states, “His experience was slender, his training had significant shortcomings, and little can be known about his thought processes and habits.” Lord Cullen found no suggestion that the driver had deliberately ignored a red signal, but thought he might have believed he had a ‘proceed aspect’. Low sunlight shining into his eyes, and the unusual configuration of the signal which impaired initial sighting of its red aspect are cited as possible explanations for his fatal misunderstanding.
It’s human nature to be horrified by such stories, and the reasons for the carnage in Galicia will continue to unfold for months to come, keeping it alive in our minds, but the fact remains that rail travel is one of the safest modes of transport out there. Besides which there’s no such thing as 100% safety. Even our own homes are potentially lethal – stairs to fall down, chip pan fires, slipping in the bath, the list goes on.
As we all go about our daily business, the quest to make trains even safer continues. But, is it simply a matter of technical innovation? Perhaps it might also involve seeking greater understanding of and integration with driver psychology and behaviour, and the way the human brain copes – or doesn’t cope – with monotony. Perhaps that’s the next big challenge for both the engineers who build the infrastructure and those who develop the train control and signalling technology?