For context, I'm a signalling industry consultant (and for many years was "the" consultant) to the Rail Safety and Standards Board specifically on EMC compatibility of track circuits, a (somewhat specific!) field I've found myself involved in for over 30 years now. So I will be reading this paper in detail with interest.

In principle, yes we get concerned about any additional sources of electrical energy that can interfere with, particularly, d.c. track circuits, as these can in theory result in an unsafe condition (wrong side failure). Modern track circuits require complex signals where it becomes increasingly unlikely that interference will be credibly detected as the "correct" track circuit signal, but d.c. track circuits only need d.c. And crucially, in the worst case of environmental conditions, they can only need a relatively small additional level of energy to trip them into false detection of a clear track. As an example, a bugbear of the industry for very many years has been the risk of battery effects, due to e.g. dissimilar metals in floodwater, which in theory can create wrong side failure of d.c. track circuits by adding a small amount of additional energy into the system.

Now, I've only very very quickly skimmed this paper over lunch (for obvious reasons I will be reading it in detail later), however my immediate thought - which may change - is that I can't see myself and my colleagues in the industry losing much sleep over this. One issue is whether this source will actually couple enough energy in to drive the detector relay - we're not talking about electronic detection as with modern track circuits, this actually has to drive what I can only describe as a hoofing great relay (although to be fair we are only talking about additional energy on top of the correct track circuit drive, so this may be credible). Far more significantly, even if this source does persist for long enough to energise the relay (they're not quick), for this situation to result in an accident it then further needs to persist long enough to allow the green signal to be illuminated, for the driver to see it, pass it, and hit the train in front. Now, in track circuit development we do consider ANY energisation of the drive relay when the train is on the track as a wrong side failure, but there is a significant time level before there is an actual railway level hazard. As far as I can see this paper only talks about voltages, rather than duration, and both are critical - but again that may be my error in reading it quickly.

To be honest what I would be (and we are) concerned about is the right side ("safe") impact. Not on d.c. track circuits, which are basically only a power supply, a resistor, and a relay and so as tough as old boots, but on modern electronic track circuits - and axle counters, and balises. These are designed to fail safe if "space weather" or any other EMC effect occurs, but "safe" means setting the signals to red - and then manually getting the train and passengers back to a place of safety. And that manual process is where accidents actually occur. So resilience to space weather absolutely is an issue, I'll let you know later what my further thoughts on the safety issues are.

But many thanks Roger for bringing this up, and for adding the link to the paper itself, I wasn't aware of this.

Thanks,

Andy

For context, I'm a signalling industry consultant (and for many years was "the" consultant) to the Rail Safety and Standards Board specifically on EMC compatibility of track circuits, a (somewhat specific!) field I've found myself involved in for over 30 years now. So I will be reading this paper in detail with interest.

In principle, yes we get concerned about any additional sources of electrical energy that can interfere with, particularly, d.c. track circuits, as these can in theory result in an unsafe condition (wrong side failure). Modern track circuits require complex signals where it becomes increasingly unlikely that interference will be credibly detected as the "correct" track circuit signal, but d.c. track circuits only need d.c. And crucially, in the worst case of environmental conditions, they can only need a relatively small additional level of energy to trip them into false detection of a clear track. As an example, a bugbear of the industry for very many years has been the risk of battery effects, due to e.g. dissimilar metals in floodwater, which in theory can create wrong side failure of d.c. track circuits by adding a small amount of additional energy into the system.

Now, I've only very very quickly skimmed this paper over lunch (for obvious reasons I will be reading it in detail later), however my immediate thought - which may change - is that I can't see myself and my colleagues in the industry losing much sleep over this. One issue is whether this source will actually couple enough energy in to drive the detector relay - we're not talking about electronic detection as with modern track circuits, this actually has to drive what I can only describe as a hoofing great relay (although to be fair we are only talking about additional energy on top of the correct track circuit drive, so this may be credible). Far more significantly, even if this source does persist for long enough to energise the relay (they're not quick), for this situation to result in an accident it then further needs to persist long enough to allow the green signal to be illuminated, for the driver to see it, pass it, and hit the train in front. Now, in track circuit development we do consider ANY energisation of the drive relay when the train is on the track as a wrong side failure, but there is a significant time level before there is an actual railway level hazard. As far as I can see this paper only talks about voltages, rather than duration, and both are critical - but again that may be my error in reading it quickly.

To be honest what I would be (and we are) concerned about is the right side ("safe") impact. Not on d.c. track circuits, which are basically only a power supply, a resistor, and a relay and so as tough as old boots, but on modern electronic track circuits - and axle counters, and balises. These are designed to fail safe if "space weather" or any other EMC effect occurs, but "safe" means setting the signals to red - and then manually getting the train and passengers back to a place of safety. And that manual process is where accidents actually occur. So resilience to space weather absolutely is an issue, I'll let you know later what my further thoughts on the safety issues are.

But many thanks Roger for bringing this up, and for adding the link to the paper itself, I wasn't aware of this.

Thanks,

Andy

P.S. I HATE the sub-heading on the E&T article "Space weather could have unintended and dangerous effects on the UK’s train networks, including switching signalling from red to green, according to new research." "Unintended" yes, "dangerous" is, at best, an "unhelpful" term. I'm sure they'd argue that they've only said "could"...which would be fine(ish) if it was in the tabloid press.

Now, leaves on the line, they genuinely do result in dangerous failures of track circuit systems, including switching signalling from red to green. If E&T want to worry their readers they could try worrying them about that...but of course that's not "news" though, it's been true for 150 years!

Rant over,

Andy

Ok, having read through it properly now: It's a good interesting paper. The authors are clear that what they've built is a coupling model, not an analysis of the risk to the railway. And equally they are clear that they have not analysed the time factor, and that this would need to be considered to determine the actual level of risk.

I think it's unfortunate that in their paper they have (I am sure inadvertently) suggested that a relay being energised leads to an unsafe condition. They've described the response time of the relay "on the millisecond scale" (it's specified in BR939A as 20ms), but that's not the issue, it's whether the fields persist long enough for that relay to stay energised such that the driver sees a green signal. It's not an edge triggered system, it's a state managed system - as soon as the relay drops back the signals will turn red / amber as appropriate. The paper isn't wrong (and the rail advisors they've quoted are good and very much know what they're doing!) but imo would have benefitted from a bit of clarity there.

So what I feel the paper completely misses is any background on how long such fields exist for in a steady state, without that it's impossible to judge with scale of a problem this is. To be fair, that's not what the authors were trying to present, all they are actually presenting is a coupling model which others can use to carry out the time varying analysis. But since presumably such data exists in the sources which the authors quote as background on the field strength (or at least in sources which are in turn related to those sources) it would have been useful to present those.

And, lest it scare anyone from travelling on the railway, d.c. track circuits have been in use in the UK for - ok I'm not actually sure (I should know!) and Google's letting me down, let's say 120 years? That's not saying this isn't a problem, it may indeed turn out to be a source of some of the odd blips we see - the pain with track circuits is you get unrepeatable blips which are impossible to pin down and could be from many causes.

So 9.99 out of 10 for this paper. But, sorry, 2 out of 10 for the E&T article. The quotes in it are literally correct, but the context is alarmist and unhelpful. All the paper has presented (and I think all it's authors intended to present) is a coupling model, with an argument to say that that model may be useful to consider a potential safety issue. That's a very long way from saying that there actually is a safety issue.

But next time we're investigating reports of d.c. track circuit blips - right side or wrong side - it's another potential cause we can throw into the pot, so to be fair to E&T they have finally published an article which might come in useful to me in the day job!  I will admit I hadn't come across this issue before, and probably should have done. Which just goes to show there's always new things to learn.

Meanwhile, when you're frustrated by train delays due to leaves on the line, be grateful that that cause of wrong side track circuit failures is being taken seriously (it doesn't do much for train braking either).

And tbh more worrying is the suggestion in this paper that the West Coast Main line really does have d.c. track circuits (and therefore block joints). Surely it's all axle counters or jointless track circuits now? Nearly 20 years' ago we were tendering to move it to ERTMS!

Thanks,

Andy

P.S. Having googled "track circuits space weather" E&T was not the worst, so far my prize for that goes to the headline "Solar storms could cause train accidents, scientists warn". Oh no, maybe it's been beaten by "Weather from outer space 'could make trains crash into each other' "

It's not a good idea to read headlines on a subject you know something about...

If it is useful the time scales of the high ampoitude events that are being worried about are minutes to hours, with the fast changing parts being sub (20 mins) cycle period, but very variable. (that would be shocks and aftershocks to the magnetosphere from the arrival of a burst of particles following a coronal mass ejection in our direction.)

From this report. looking at the Spanish grid.

There are however other faster mechanisms, usually lower amplitude and also there is evidence that the near-equatorial events contain more high frequency components

From this paper

Equally there are slower patterns that are essentially related to the earths rotation  period, i.e. 24 hours - if you wait long enough, you will see it....

A longer read, with some explanations of the mechanisms, in this thesis.

Mike.

Thank you Andy,

The E&T headline and article triggered my BS detector. The actual paper seemed reasonable but I don't have the background knowledge to judge it properly.

The headlines very often don't match the actual science/engineering.

Thanks Mike, the interesting question for this issue is how those disturbances periods relate to d.c. shifts, it looks as if the underlying data may all be available (e.g. in that very useful thesis you've linked to), but it may need to be modelled differently to determine the expected periods of net d.c. voltage, and then an overall model built, incorporating the model described in the Lancaster paper, to determine the actual risk.

I won't comment here on how this may or should happen as I'm not going to risk treading on toes. But generically it's that interesting conundrum: to determine if there is actually a credible risk requires funding, but how do you justify allocating funding when you don't know if there is a credible risk!

Cheers,

Andy

Ah - I suspect the sensational if off the track headlines are the mechanism to drive the availability of funds, first  for a 50 page 2 week case-study by some well known consultancy or other, probably at about a grand a page, for the civil service to digest,  and then when this concludes, as it must of course, as all you can do in 2 weeks is a literature survey, that further experimental work is needed as the risks are as yet unknown and hard to model, then the university gets what it always wanted, research funding for another student for a 3 year PhD.....

Or I may be a slightly too cynical so-and-so...

Mike.

I wouldn't have dared write that (particularly in this case), but...yeeeeah.

(Incidentally my time for Track Circuit EMC consultancy, at least to RSSB, is given by my employer pro bono, which is actually a bit odd.)

That said, no-one, not even a university, is going to make a fortune from rail industry research! I guess for this issue I best fit into the "civil service" part of your description, and we simply don't have access to funding to fund projects that last more than three months at a time. Which in turn brings its own problems. 

Generally issues like this would actually have to be funded by suppliers, on the principle that it's the supplier's problem to prove that their system is safe, unfortunately in this case there is no supplier (at least none with the technical knowledge to look at this) as d.c. track circuits are basically obsolete - one of my jobs when I worked for one the major suppliers was to make our version of them obsolete!