01.03.15
Rail safety performance and trends
Source: Rail Technology Magazine Feb/March 2015
The RSSB has released its high-level Overview of Safety Performance for the calendar year 2014. RTM discussed the results with George Bearfield, director of system safety at RSSB.
Overall 2014 was a good year for railway safety, and it was the seventh straight year where there were zero deaths in train accidents.
In some areas, however, there was some less good news, and on a number of different metrics it seems as if railway safety ‘peaked’ in the early years of Control Period 4, around 2010-12, and has since fallen back a little.
This includes the overall number of fatalities (excluding trespass and suicide), the number of SPADs (signals passed at danger), and the number of PHRTAs (potentially higher-risk train accidents).
Trends
Asked about these trends, particularly in fatalities, George Bearfield, RSSB director of system safety, told us there is a degree of statistical uncertainty. He said: “It’s always difficult when you look at an absolute measure like this one; you have to overlay against this the fact that journeys and passenger numbers are increasing, and have been increasing historically. Our normalised charts are better to look at in terms of a comparison of how the industry’s performing. The more trains you have, the more people you have, the more opportunities there are for an incident.
“It’s been recognised in the industry that passenger safety is a particular focus, so we recently launched the new platform train interface strategy, which is about the whole industry coming together and coming up with a joined up strategy about how we address risk at the platform train interface.” (More on page 84)
Bearfield continued: “It is difficult to draw specific conclusions from these top-level figures. The reason we monitor this, and why we work through our various stakeholder groups, is when we do see changes we ask questions, and that drives how we organise as an industry and where we put our focus into trying to bring safety improvements.
“We’re certainly not complacent. In terms of long-term safety performance, we did a lot of work analysing how the industry performed over CP4 and we were satisfied that there were reductions in line what the DfT was looking for, in terms of passenger and workforce safety.”
Signals passed at danger
At 292, the number of SPADs in 2014 was 2% higher than the 285 in 2013. At the end of 2014, SPAD risk stood at 61% of the September 2006 baseline level, down from 70% at the end of 2013. We asked Bearfield to explain the disparity between risk and overall numbers, and he said: “Risk is a measure of fatalities and weighted injuries. Obviously we don’t get many, if any, visible injuries as a result of SPADs. So we have to try to work out where the risk level is by doing more analysis.”
The SPAD Risk Ranking Methodology dives deep into the circumstances and potential consequences of each SPAD, which ranks them on a logarithmic scale.
He said: “One SPAD is not as bad as another – if you pass a signal at danger and go past the conflict point where there would have been the potential for a collision had another train happened to have been there, you’re down to chance. That’s a ‘worse’ SPAD in terms of our risk ranking than one where if the train travelled at low speed and came to a stop well short of the conflict point. Understanding that difference in potential risk is very important in terms of making sure we’re investing efficiently in the railway to improve safety.”
A SPAD risk of zero is impossible, Bearfield said – “there is always something, no matter how infinitesimally small, that could go wrong”. But ERTMS, an “inherently more failsafe method of signalling and of train operation”, should help cut SPAD risk, as should further research on low adhesion.
Reduction in public risk
The RSSB also uses a sophisticated ‘precursor indicator model’ (PIM) to extrapolate potential risk from data like broken rails and other useful indicators.
Bearfield said: “The Safety Risk Model is only updated every couple of years – the PIM uses certain ongoing indicators to scale that risk up and down.”
The report notes: “At the end of 2014, the underlying passenger risk from train accidents was 3.02 FWI (fatalities and weighted injuries), compared with 3.14 FWI at the end of 2013, which is a 3.9% decrease. The overall PIM at the end of 2014 was 6.83 FWI, compared with 7.61 FWI at the end of 2013, which is a decrease of 10.3%. The improvement in overall PIM risk is mainly due to a reduction in public risk.”
Asked about that reduction, Bearfield said: “Because it’s an industry-wide metric, it’s difficult to diagnose exactly what is causing increases and decreases in many areas.”
He said major changes like the installation of TPWS (train protection and warning system) drove clear reductions in SPAD risk, but with overall risk now at a much lower level, “it is difficult to attribute particular causes”.
The main Safety Risk Model is now being upgraded, linking in with the more granular data being produced by things like Network Rail’s ORBIS (Offering Rail Better Information Services) programme.
Trusting the stats
Bearfield said he is “very confident” in the statistics in the report, because of the huge amount of effort put into data quality and data health checks.
But he added: “There can be a danger in a short-term reaction to something, and differentiating between something statistically not significant because of the randomness in the data and something which is more meaningful. Ultimately, that comes down to more than just data: it comes down to investigating the underlying issues in more detail, which the industry and RSSB as a whole can do and does do. They’re as robust as we can make them at present and they give a really good view. But they always have to be complemented with expertise and understanding and further investigation.”
Discussing the way that rail industry risk management models and research have been taken up by other industries, Bearfield told us that aviation, oil & gas and the nuclear sector have all been involved. The Highways Agency risk model was based on rail’s, Bearfield said, adding: “With the Highways Agency coming under the ORR’s remit, I think we’re going to have a lot more scope and opportunity to not just share good practice with them but to do more integrated analysis – to actually work together in understanding risk. There are some shared issues.”
Road risk
The road network is in fact a direct safety risk for the rail industry too – members of the workforce are more likely to die on the roads between or after jobs than while on the railway infrastructure. RTM reported in June 2014, for example, on the crash that killed track workers Steven Sheldon, Martin Williams, and James Vernon Stark, who were on their way home to south Wales after a shift in Reading. Such ‘off-duty’ deaths are not included in the RSSB’s Overview analysis, though its scope is being reviewed. The newest Overview does include the death on 1 May of BAM worker Huw Jenkins, from Port Talbot, Wales, who was involved in a road traffic accident on the A7 north of Galashiels while working on the Borders Railway project.
Bearfield said the System Safety Risk Group had “flagged” road driving risk as a priority area. “It can be a hidden risk – it’s not always understood,” he said. “Historically, it’s been difficult to gather the data we need to analyse it properly. Obviously there’s a huge contractor workforce on the rail network and – as the report points out – in recent years there has been a number of road traffic accidents where members of the railway workforce has been injured. I think the industry has rightly identified that as an area of focus.”
More detailed safety analyses will be in the RSSB’s Annual Safety Performance Report, the next of which will be published at the end of June 2015.
Headlines from the report
For the seventh year in succession, there were no passenger or workforce fatalities in train accidents.
- Excluding trespass and suicide, the total number of fatalities in 2014 was 17, compared with 18 in 2013.
- Three of the fatalities were passengers, compared with six in 2013; all occurred at stations. Three members of the workforce were fatally injured in 2014; two were fatally injured in 2013. Excluding trespass and suicide, 11 members of the public were fatally injured, compared with 10 in 2013.
- The number of potentially higher-risk train accidents (PHRTAs) in 2014 was 32, compared with 30 occurring in 2013.
- At 292, the number of signals passed at danger (SPADs) in 2014 was a 2% increase on the 285 recorded in 2013. SPAD risk rose to 84% mid-year, before reducing and ending 2014 at 61% of the September 2006 baseline level. This compares with 70% at the end of 2013.
- Fatalities arising from trespass and suicide totalled 320 in 2014, compared with 315 in 2013. 2014 was the third consecutive year the total number increased.
- All this information should be treated as provisional.
Tell us what you think – have your say below or email [email protected]