Ely Upgrade under threat

[Starmill]

this is the worst bit, number of trains per day measures the average risk, but what's required is to measure the peak risk, which in most cases will be at the times of day when there is most traffic, both on the rails and on the roads.

But this is also fundamentally an incorrect way of looking at it.

If you double the number of train movements you double the number of opportunities for a rail vehicle to collide with an obstruction on the crossing.

All of this is why a full understanding of level crossing safety needs to consider the number of people using it on trains, the number of people in vehicles or on foot or bicycle, how many such movements there are and all of the circumstances around each movement.

 

[Taunton]

Actual experience, as opposed to arithmetic theory, shows actual incidents are significantly dependent on oddball operating issues rather than regular arrangements.

Athelney showed that closing the AHB the best part of 10 minutes before the train, instead of the normal short interval, due to a combination of a tamper having been run wrong line over the treadles, and the signal centre not knowing what this did to the sequence, caused a motorist who lived within sight of the crossing and used it multiple times a day to assume it had failed.

One on the Southern on a double track line with a half-hourly service had a train one way, then staying down for an opposite direction one, then staying down again for an unusual movement following the first had a motorist misled into thinking they had failed - how can there be three trains on a quiet two-track line.

My own experience at a failed AHB (and failed phones), in a storm, where service had been suspended as well, had me out checking the line clear, then instructing my passenger explicitly to stand and look both ways on the line as I drove over, had them do something else as I did so - people are just stupid.

All these covered here before by me so no need to repeat. But none of these situations is envisaged in any risk assessment.

 

[Starmill]

Oh, I expect you're correct. And this perfectly elaborates what I said in the previous post - this sums up a great deal of what is currently wrong with this country.

I don't think personally that it's a failing of the country that the public demands a level of safety assurance from transport providers. Although I do sometimes think that it is a failing that we see private transport being held to shockingly low standards on occasions.

 

[Magdalia]

It appears from your posts however that you don't have a subject specific understanding of how crossing safety can be assessed. As a result your grand pronouncements about litigation are very likely to be wrong.

The safety of the crossing will change with any changes in usage. This is categorised over time, but any change the industry wants to make will influence the way the crossing is categorised. This will require the modelling to be done again.

If the number of people on the train rises or the number of vehicles crossing the railway at the crossing rises, the risk will increase and thus mitigations will be recommended by the review of the crossings.

The mitigations are costly and sometimes difficult to implement. Sometimes no mitigation can be recommended because the only way to keep risk within acceptable levels is to close the crossing.

It is reassuring to read that these are reviewed, though I am still concerned that peak risks are not being mitigated and possibly misallocation of resources because of use of daily rather than peak movements. And, despite what you say, I see no recognition that train movements are not independent, and how that changes the risks.

But this is also fundamentally an incorrect way of looking at it.

If you double the number of train movements you double the number of opportunities for a rail vehicle to collide with an obstruction on the crossing.

That's very simplistic, taking no account of the fact that train movements are not statistically independent. And the doubling of a tiny risk is still a tiny risk: on its own it is very unlikely to require intervention.

 

[Magdalen Road]

Hmm. I respectfully suggest the survey is wrong. 4,400 vehicles is an average of once every 20 seconds, day and night. Given there's hardly any traffic at night, it suggests a vehicle every few seconds during the day. That is just plain wrong as anybody who knows the area will instantly realise.
If these are the numbers feeding into the crossing risk assessment, it's conclusions are also wrong.

There were cameras and cables over the roads, at several junctions as they were monitoring where the traffic was going to / from.
I saw this with my own eyes.
It's been awhile since I read the report but there is a surprising amount of traffic using the roads round there, as I explained earlier. The A142 is busy and dangerous, traffic still backs up from Ely to Stuntney even with the new bypass.

 

[Starmill]

It is reassuring to read that these are reviewed, though I am still concerned that peak risks are not being mitigated and possibly misallocation of resources because of use of daily rather than peak movements. And, despite what you say, I see no recognition that train movements are not independent, and how that changes the risks.

I don't think anyone has suggested that the risk at what you describe as peak time isn't higher. It is higher when more people are on the train and when more trains run. It is also higher when more road vehicles or pedestrians are trying to use the crossing around the same time. The model, however, will account for all of this. What you were saying was that overall, risk doesn't change based on there being a larger or smaller difference between peak and off peak traffic. It was only in that which you were wrong.

That's very simplistic, taking no account of the fact that train movements are not statistically independent. And the doubling of a tiny risk is still a tiny risk: on its own it is very unlikely to require intervention.

It's not remotely simplistic. I was pointing out that you were claiming to understand risk management, but had ignored some of the critical elements from which risk arises, i.e. the total number of train movements and the number of people on said trains.

For railways, is almost as important to understand the likely consequences of a collision as the risk of any collision at all taking place.

 

[stuu]

Hmm. I respectfully suggest the survey is wrong. 4,400 vehicles is an average of once every 20 seconds, day and night. Given there's hardly any traffic at night, it suggests a vehicle every few seconds during the day. That is just plain wrong as anybody who knows the area will instantly realise.
If these are the numbers feeding into the crossing risk assessment, it's conclusions are also wrong.

The DfT traffic counts page shows a traffic count site east of Queen Adelaide. The data ends in 2009, but the highest value was 2900 vehicles per day. Add in the traffic from the village itself, which logically is mostly going to Ely, and a general increase in traffic over 13 years, and 4400 doesn't sound implausible

 

[MikeWM]

I don't think personally that it's a failing of the country that the public demands a level of safety assurance from transport providers. Although I do sometimes think that it is a failing that we see private transport being held to shockingly low standards on occasions.

I entirely agree, but I equally think it is a significant failing if we demand an *excessive* level of safety assurance, particularly if that results in a situation where we can't realise the large benefits of a necessary infrastructure upgrade, which is exactly what we appear to be doing here.

 

[Maltazer]

Hmm. I respectfully suggest the survey is wrong. 4,400 vehicles is an average of once every 20 seconds, day and night. Given there's hardly any traffic at night, it suggests a vehicle every few seconds during the day. That is just plain wrong as anybody who knows the area will instantly realise.
If these are the numbers feeding into the crossing risk assessment, it's conclusions are also wrong.

Don't forget that's traffic in both directions, so half that each way. Most of the traffic will be in the rush hours too.

Here's a link to the survey
https://*******.com/2s3e4nkr

And another to the risk assessment from 2015 for the Kings Lynn line crossing - it quotes traffic surveys from 2013 and 2015 that show similar numbers to the 2016 one.
https://*******.com/a2nsnyjc

Three separate surveys performed in different years are not likely to all come up with similar numbers but somehow all be wrong.

Edit - don't know why links not working. They're tiny url

 

[Mordac]

That’s what I said!





because, if it is accepted ‘in this one case’, then others will want the same in the same circumstances, and one case becomes many.

Also, one day, there will be an incident at the crossing. And someone who signed off the change will have to stand in front of the judge to answer the question from the prosecution “the risk assessment showed that the risk assessment of an incident increased by x, and you chose not to implement the proposed mitigations That would reduce that. As the controlling mind, that makes it your fault. How do you plead To the charge of manslaughter by gross negligence?”

if you think I’m scaremongering, I’m not. I know people who have been through this (or the precursors to it).

Wouldn't the person doing that be a Network Rail employee, and as such an agent of the state with sovereign immunity?

 

[Starmill]

Wouldn't the person doing that be a Network Rail employee, and as such an agent of the state with sovereign immunity?

I don't think that's a relevant defence to charges of corporate manslaughter. It's likely that Network Rail would qualify as a defendant.

 

[Bald Rick]

Indeed, and assessing LX risk related to the total number of trains over a crossing per day is fundamentally flawed in three different ways.

• Each train passing over a level crossing is not of equal risk: for example length of train, speed and time of day will all lead to different amounts of risk;
• Train movements are not independent, because they run to a timetable, and a train movement following closely after another movement will have a different risk to a train where there has been a long gap since the previous movement: this means that risk cannot be aggregated by summing the risk of each individual movement;
• this is the worst bit, number of trains per day measures the average risk, but what's required is to measure the peak risk, which in most cases will be at the times of day when there is most traffic, both on the rails and on the roads.

Applying these to Ely-Kings Lynn:

• the Middleton sand trains will be of different risk to the Kings Cross passenger trains;
• the risk at Littleport bypass has increased significantly since the 2018 timetable change, even though the number of trains barely changed, because off peak the trains now pass at Littleport not Downham;
• the risk should be managed at 4 trains per hour already, because that is the peak risk.

Indeed again. If a litigant's solicitor called a risk management professional to give evidence then Network Rail's risk assessments based on number of trains in a day would not stand up in court.

basically what you’re saying is that the risk assessment methodology developed by Network Rail in conjunction with RSSB, and approved by ORR, and developed over the past 15 years through further empirical evidence, which is used by ORR to hold NR to account is wrong, but your methodology is right?

I’ll leave it to the forum readership to judge.

Where else in the country does a road cross 3 level crossing on 3 different rail lines in little more than a quarter of a mile? I'm struggling to think of another example. Ely North/Prickwillow Road is a unique circumstance.
Even if that were not the case, common sense would say if alternative mitigations are acceptable at Ely, they may well be acceptable elsewhere.

The location and local circumstances are largely irrelevant. Each Lx has to be considered independently, and it is the change in risk that’s matters.

Wouldn't the person doing that be a Network Rail employee, and as such an agent of the state with sovereign immunity?

No, and I‘ve never heard of such a concept in this context.

 

[trebor79]

The location and local circumstances are largely irrelevant. Each Lx has to be considered independently, and it is the change in risk that’s matters.

If each crossing is considered independently, why would recognising the unique situation an Queen Adelaide and doing something that costs less than £450m mean there would be pressure to reduce standards elsewhere?
It seems that the approach being used does not consider each crossing independently, and merely feeds data about traffic levels into a model than churns out a risk rating and therefore what is deemed to be an acceptable solution. That is the antithesis of treating each individually and taking into account unique factors that no data driven model can account for.

If we accept £450m is an absurd figure and isn't going to happen, what's the alternative other than do nothing and have an ever more restrictive bottleneck?
There must be something that can be done. Mankind does all kinds of amazing things, to say we can't come up with an economically sensible solution to 3 level crossings accessing a hamlet and a haulage yard is ridiculous.

 

[Bald Rick]

The £450m is not just for sorting those crossings, it’s for a lot, lot more!

 

[trebor79]

The £450m is not just for sorting those crossings, it’s for a lot, lot more!

Yes I was forgetting that. How much of it was for sorting the crossings, out of interest? I'm guessing quite a significant proportion.

 

[Sonik]

I entirely agree, but I equally think it is a significant failing if we demand an *excessive* level of safety assurance, particularly if that results in a situation where we can't realise the large benefits of a necessary infrastructure upgrade, which is exactly what we appear to be doing here.

The main thing to be borne in mind IMO, is not the perceived level of risk (which by definition contains subjective components) but the value of the improvement, because every pound spent is a pound that is not available to spend on safety improvements elsewhere.

 

[bspahh]

Don't forget that's traffic in both directions, so half that each way. Most of the traffic will be in the rush hours too.

Here's a link to the survey
https://*******.com/2s3e4nkr

And another to the risk assessment from 2015 for the Kings Lynn line crossing - it quotes traffic surveys from 2013 and 2015 that show similar numbers to the 2016 one.
https://*******.com/a2nsnyjc

Three separate surveys performed in different years are not likely to all come up with similar numbers but somehow all be wrong.

Edit - don't know why links not working. They're tiny url

This table is on page 20 of the Queen Adelaide Level Crossing Traffic Survey from January 2018

 

[Magdalen Road]

If each crossing is considered independently, why would recognising the unique situation an Queen Adelaide and doing something that costs less than £450m mean there would be pressure to reduce standards elsewhere?
It seems that the approach being used does not consider each crossing independently, and merely feeds data about traffic levels into a model than churns out a risk rating and therefore what is deemed to be an acceptable solution. That is the antithesis of treating each individually and taking into account unique factors that no data driven model can account for.

If we accept £450m is an absurd figure and isn't going to happen, what's the alternative other than do nothing and have an ever more restrictive bottleneck?
There must be something that can be done. Mankind does all kinds of amazing things, to say we can't come up with an economically sensible solution to 3 level crossings accessing a hamlet and a haulage yard is ridiculous.

As previously mentioned it's not just a hamlet - it's the only route in / out of east Ely, links to Littleport, provides a safe alternative route to the A142. The A road is frequently blocked due to accidents, lots of lorries and tractors.
Several new estates appeared since 2009. The Thistle Corner junction was opened, so the estate (pink arrow) now uses the Prickwillow road as does much other traffic (roundabout to the right). The red arrows indicate an estate that's being built behind that, which now also joins Thistle Corner. The left red arrow is a large estate being built, around 10,000 people iirc. Some of the traffic will use the A10 but that will also increase traffic through Queen Adelaide.
Ely is expanding so the road activity through the level crossing area is likely to go up significantly.

 

[yorksrob]

Most risk at level crossings is as a result of by misuse by motorists, therefore the mitigations should be paid for out of the highway budget, not the rail budget.

 

[ac6000cw]

Most risk at level crossings is as a result of by misuse by motorists, therefore the mitigations should be paid for out of the highway budget, not the rail budget.

The funding comes from the same source in the end for both rail and road - taxation and government borrowing.

The £450m is not just for sorting those crossings, it’s for a lot, lot more!

Exactly! (and something I've pointed out on this thread as well) - also I think that figure was just an NR estimate before the consultation phase started.

As previously mentioned it's not just a hamlet - it's the only route in / out of east Ely, links to Littleport, provides a safe alternative route to the A142. The A road is frequently blocked due to accidents, lots of lorries and tractors.
Several new estates appeared since 2009. The Thistle Corner junction was opened, so the estate (pink arrow) now uses the Prickwillow road as does much other traffic (roundabout to the right). The red arrows indicate an estate that's being built behind that, which now also joins Thistle Corner. The left red arrow is a large estate being built, around 10,000 people iirc. Some of the traffic will use the A10 but that will also increase traffic through Queen Adelaide.
Ely is expanding so the road activity through the level crossing area is likely to go up significantly.

Well said!

As well as a rail hub, Ely is a major nodal point on the Fenland road network, which is very important to the local economy, particularly for agricultural traffic. The road that runs over the crossings at Queen Adelaide is effectively part of a de-facto Ely eastern bypass - just get off a train in Ely, turn right out of the station and take a walk through the underpass to the junction where the 'yellow' road in the map above heads towards Queen Adelaide and see how much traffic goes to and from that direction.

 

[yorksrob]

The funding comes from the same source in the end for both rail and road - taxation and government borrowing.

That might be true for the government, but it will have a very tangible impact on the department/service whose budget it ends up depleting.

 

[MikeWM]

The main thing to be borne in mind IMO, is not the perceived level of risk (which by definition contains subjective components) but the value of the improvement, because every pound spent is a pound that is not available to spend on safety improvements elsewhere.

I don't think the value of improving Ely North is in any real doubt, is it? It has cropped up on pretty much any list of 'most important things on the network to sort out' for a good number of years.

 

[Sonik]

I don't think the value of improving Ely North is in any real doubt, is it? It has cropped up on pretty much any list of 'most important things on the network to sort out' for a good number of years.

It's the value of the proposed 'improvements' that are debatable.

 

[Magdalia]

It appears from your posts however that you don't have a subject specific understanding of how crossing safety can be assessed. As a result your grand pronouncements about litigation are very likely to be wrong.

Subject specific understanding is key to the identification of the risks at each location, and the design of the mitigations; I've tried not to stray into that territory.

But the principles of risk management are universal. The railway is not an exception to that.

On the "grand pronouncement", I take your point!

For railways, is almost as important to understand the likely consequences of a collision as the risk of any collision at all taking place.

Most risk at level crossings is as a result of by misuse by motorists

Absolutely. I have been involved in risk management of competitive sporting events on public roads not closed to traffic. Collisions are high impact risks, literally. The risk assessment would be based on level of road traffic and location specific features. The number of competitors would be irrelevant, except at the start and finish areas, unless it was a "mass event". The notion that a collision risk could be identified, and that we would decide that mitigation was not required with 52 competitors, but was required with 70 competitors, would have been regarded as ridiculous.

basically what you’re saying is that the risk assessment methodology developed by Network Rail in conjunction with RSSB, and approved by ORR, and developed over the past 15 years through further empirical evidence, which is used by ORR to hold NR to account is wrong

There's a lot of railway industry insiders there, but does it include any independent peer group appraisal from outside the industry, or benchmarking against different industries?

 

[MikeWM]

It's the value of the proposed 'improvements' that are debatable.

Well, I did question a couple of pages back if they were sufficient or whether even more capacity is likely to be needed soon

But something clearly needs to be done at Ely North, even if all the other stuff gets kicked into the long grass.

 

[InOban]

Most risk at level crossings is as a result of by misuse by motorists, therefore the mitigations should be paid for out of the highway budget, not the rail budget.

Surely the housing developers should be paying?

 

[yorksrob]

Surely the housing developers should be paying?

I suppose in this case it depends on whether the upgrade is for the village, or the crossing is used as a main thoroughfare to the North East !

I don't have knowledge of such things, but level crossings are one instance where the railway (and consequently passengers) have to carry the can for bad road driving.

 

[Bald Rick]

but does it include any independent peer group appraisal from outside the industry

yes.

 

[Sonik]

So..about Ely North!

 

[Bikeman78]

That’s what I said!





because, if it is accepted ‘in this one case’, then others will want the same in the same circumstances, and one case becomes many.

Also, one day, there will be an incident at the crossing. And someone who signed off the change will have to stand in front of the judge to answer the question from the prosecution “the risk assessment showed that the risk assessment of an incident increased by x, and you chose not to implement the proposed mitigations That would reduce that. As the controlling mind, that makes it your fault. How do you plead To the charge of manslaughter by gross negligence?”

if you think I’m scaremongering, I’m not. I know people who have been through this (or the precursors to it).

Does this only apply to railways? If not then I look forward to someone in Cardiff Council going to prison when a cyclist gets mowed down on the new cross city lane. It will happen sooner or later. It's already nearly happened to me.

You can't mitigate by putting in full barriers with obstacle detection, because that leads to the crossing beings closed to traffic for an unacceptable amount of time.

The only way to avoid expensive interventions is to increase the number of trains and just leave the crossings as half barriers - and it doesn't sound like anyone would sign off on that.

Maybe they should spend a week in the Netherlands. Even I raise an eyebrow over there. It's fascinating watching cars streaming across crossings with the train already in view travelling at 140 km/h.