What technology could be used to check if a line is clear?

[Winthorpe]

Moderator note - split from

Storm Darragh - 6/7th Dec

Amber wind warnings now issued by the Met Office, would expect some disruption if it plays out as forecast. Fortunately it doesn't look too wet, apart from in the welsh highlands. Gusts of 70 to 80 mph are likely around exposed coasts and headlands, where some very large waves are likely...

www.railforums.co.uk

The lines do need to be checked. I wonder if some use of modern technology could be considered. I’m thinking a drone following the line to check for obstructions. With AI to detect potential obstacles. A drone version of the New Measurement Train.

Obviously it would a first once over, with the first train proving the route properly if no obstruction was found.

Just a thought.

 

[aaronspence]

The lines do need to be checked. I wonder if some use of modern technology could be considered. I’m thinking a drone following the line to check for obstructions. With AI to detect potential obstacles. A drone version of the New Measurement Train.

Obviously it would a first once over, with the first train proving the route properly if no obstruction was found.

Just a thought.

They already have helicopters for this whch are far better suited, longer duration etc than a drone.

 

[Winthorpe]

They already have helicopters for this whch are far better suited, longer duration etc than a drone.

They do an excellent job. But can’t cover the whole railway after a storm.

A few drones in strategic locations to survey routes would be beneficial.

As I say it might be a non-starter. Just something to think about.

 

[MotCO]

Thinking completely out of the box. Ideally, an electric pulse could be sent down the rails, and if there is something on the lines, it would short-circuit. However, it assumes the obstruction can conduct electricity which is unlikely. Could sonar be used instead? Send a noise down the rail, and if it doesn't 'ping' correctly, there could be something resting on it.

 

[Darandio]

Could sonar be used instead? Send a noise down the rail, and if it doesn't 'ping' correctly, there could be something resting on it.

Not sure how that could work with things like expansion joints, points etc.

 

[30907]

Thinking completely out of the box. Ideally, an electric pulse could be sent down the rails, and if there is something on the lines, it would short-circuit. However, it assumes the obstruction can conduct electricity which is unlikely. Could sonar be used instead? Send a noise down the rail, and if it doesn't 'ping' correctly, there could be something resting on it.

And if the tree is blocking the route but not touching the rails?

 

[The Planner]

NR does have a drone team.

 

[Winthorpe]

Aerial maps are produced by constant drone surveys.

I’m just suggesting it could be used to survey disrupted routes after storms.

At least it could better direct NR crews to points where the rail needs to be cleared.

 

[aaronspence]

Surely the best method is just sending a team on a train at a very slow speed? After all, if any tech detects a problem at location X, you have to get a team there anyway.

 

[Ghostbus]

You're only going to get a serious benefit (saved time, money, resources) if you can get drones to fly inspection routes automatically, and the AI can reliably detect and ideally grade potential issues in wet and windy conditions. And that's just the technological barriers, but it's probably all the other barriers to innovation (finance, culture, regulatory) that will be harder to overcome.

 

[Indigo Soup]

If you wanted something remote, some kind of CCTV system would seem perfectly feasible.

 

[stuu]

Thinking completely out of the box. Ideally, an electric pulse could be sent down the rails, and if there is something on the lines, it would short-circuit. However, it assumes the obstruction can conduct electricity which is unlikely. Could sonar be used instead? Send a noise down the rail, and if it doesn't 'ping' correctly, there could be something resting on it.

Presumably track circuits already would do this if there were something metal across the lines?

Wood, not so much

 

[Llanigraham]

Presumably track circuits already would do this if there were something metal across the lines?

Wood, not so much

And they cannot sense something that is fouling the loading gauge, like hanging trees.

 

[zwk500]

Thinking completely out of the box. Ideally, an electric pulse could be sent down the rails, and if there is something on the lines, it would short-circuit. However, it assumes the obstruction can conduct electricity which is unlikely. Could sonar be used instead? Send a noise down the rail, and if it doesn't 'ping' correctly, there could be something resting on it.

Only works if an obstruction is directly contacting the track, which is often not the case. Also the technicals of setting up a sonar system on the rails are very complex.

The obvious thing I can think of is either to have a drone pilot that scans the line ahead a few stations, verifies it's clear, then travels on the train to the limit of the previous scan, rinse and repeat. Any problems can instantly be geo-flagged to the drone images to mobilise the response accordingly. OK you have to wait at each stop a bit longer, but does mean the driver can then be confident to proceed at maximum permitted speed, which may well be a better balance than just running at caution throughout.
Or have a front-facing LIDAR/Laser scanning box that slots onto the front of a loco/unit (where a headboard is usually placed) and will produce an audible warning if it sense an anomaly in the profile ahead. NR already has the route profiles scanned by the Gauging Train so with a bit of thought and effort a scan against that database is plausible. You'd need a few of them at each depot so trains could quickly scan routes ahead of the main service trains.

 

[Indigo Soup]

Or have a front-facing LIDAR/Laser scanning box that slots onto the front of a loco/unit (where a headboard is usually placed) and will produce an audible warning if it sense an anomaly in the profile ahead.

Could one not put such a system inside the cab? Maybe we could call it Dual-Role Inspection & Visual Emergency Response?

 

[MotCO]

Or have a front-facing LIDAR/Laser scanning box that slots onto the front of a loco/unit (where a headboard is usually placed) and will produce an audible warning if it sense an anomaly in the profile ahead. NR already has the route profiles scanned by the Gauging Train so with a bit of thought and effort a scan against that database is plausible. You'd need a few of them at each depot so trains could quickly scan routes ahead of the main service trains.

I like the idea of scanning the line ahead compared with a previously taken image. This could presumably also work in the dark with infrared?

 

[zwk500]

Could one not put such a system inside the cab? Maybe we could call it Dual-Role Inspection & Visual Emergency Response?

The advantage of the scanning box is that it would be able to automatically send back an instant image to a central system with a geographic tag and timestamp that can be used to help plan the response. It would also be able to flag everything, rather than the driver stopping every time there's anything that *might* need reporting.

It would not replace the driver conducting their own checks and reporting if needed.

 

[TurboMan]

Surely the best method is just sending a team on a train at a very slow speed? After all, if any tech detects a problem at location X, you have to get a team there anyway.

Yes, that's why there's a new section in rule book module TW1 on route proving trains.

 

[MalaiseForever]

We already use Drones in Central Route for this sort of thing.

 

[Winthorpe]

The obvious thing I can think of is either to have a drone pilot that scans the line ahead a few stations, verifies it's clear, then travels on the train to the limit of the previous scan, rinse and repeat. Any problems can instantly be geo-flagged to the drone images to mobilise the response accordingly. OK you have to wait at each stop a bit longer, but does mean the driver can then be confident to proceed at maximum permitted speed, which may well be a better balance than just running at caution throughout.

I had in mind fixed wing commercial drones. Not a small one with a radio control.

The New Measurement Train films the track. The footage is used to detect faults with AI software. It recognises patterns.

I had in mind something similar, but from above to check the track. It would be a triage process to direct crews to right people to the right place.

Trees need to be cleared in these places so send the right people there. Flood damage is in these places, send a crew there to deal with that. And so on.

If you wanted something remote, some kind of CCTV system would seem perfectly feasible.

The drone is basically a mobile CCTV.

Drones are operating every day to provide and update the aerial maps we can see on Google Maps and the like.

 

[Meerkat]

Route proving autonomous trolleys. If it doesn't have to fly then it is less affected by weather and has much less power/range restriction.
Either fancy lidar or just a loading gauge hoop, or just a camera. As its unmanned and light (and rubber tyred?) it can go much faster to the first obstruction.
Make it modular so that it can be quickly put together anywhere (what is the DEI synonym for manpackable?), and when the clearance crew get to the first obstruction they can lift it over/through and set it off to find the next one.
I'm thinking three wheeled modular/folding skeleton frame, a sensor/ control box with a handle, a portable battery/generator.
Is there a way of activating track circuits with such a light vehicle, and how do level crossing treadles work?

 

[Technologist]

Route proving autonomous trolleys. If it doesn't have to fly then it is less affected by weather and has much less power/range restriction.
Either fancy lidar or just a loading gauge hoop, or just a camera. As its unmanned and light (and rubber tyred?) it can go much faster to the first obstruction.
Make it modular so that it can be quickly put together anywhere (what is the DEI synonym for manpackable?), and when the clearance crew get to the first obstruction they can lift it over/through and set it off to find the next one.
I'm thinking three wheeled modular/folding skeleton frame, a sensor/ control box with a handle, a portable battery/generator.
Is there a way of activating track circuits with such a light vehicle, and how do level crossing treadles work?

For British Railway distances a fixed wing drone would be perfectly adequate as far as range and sensor suit. It would also be much faster, I would assume such a system would pay for itself in additional revenue form more rapidly opened lines very rapidly.

It's also possible to buy tilt rotor drones which could take off vertically but cruise efficiently. They can also change mode and enter tunnels at low speed.

 

[Meerkat]

For British Railway distances a fixed wing drone would be perfectly adequate as far as range and sensor suit. It would also be much faster, I would assume such a system would pay for itself in additional revenue form more rapidly opened lines very rapidly.

It's also possible to buy tilt rotor drones which could take off vertically but cruise efficiently. They can also change mode and enter tunnels at low speed.

Flying drones have restrictions on weather and range, and if they get bigger then approvals is going to start being an issue.
I thought wheeled because they can see better from rail level rather than trying to look down and through tree cover, and if your drone is low enough to go through tunnels it will be an issue with OHLE and will need line Block co-ordination.

 

[Winthorpe]

Flying drones have restrictions on weather and range, and if they get bigger then approvals is going to start being an issue.
I thought wheeled because they can see better from rail level rather than trying to look down and through tree cover, and if your drone is low enough to go through tunnels it will be an issue with OHLE and will need line Block co-ordination.

My suggestion to use drones was to gather data after these events. To work out what potential problems there are, if any, affecting the route. And direct the appropriate resources to the affected locations.

The route will still have to be proved. But the process is speeded up if, for example, the drone survey shows there are three trees down at x, y, z locations.

Without the aerial survey the route proving team is in the dark until they find the problem on the ground.

Technology like this:



This was done by a manned aircraft. But large commercial drones do this too.

The photo is a screenshot from the FlightRadar24 blog. See here.

 

[Llanigraham]

My suggestion to use drones was to gather data after these events. To work out what potential problems there are, if any, affecting the route. And direct the appropriate resources to the affected locations.

The route will still have to be proved. But the process is speeded up if, for example, the drone survey shows there are three trees down at x, y, z locations.

Without the aerial survey the route proving team is in the dark until they find the problem on the ground.

And where are the repair teams going to be based whilst the drone is flying around?
At least at the moment as soon as the proving train gets to an obstruction the team on-board get off and try and sort it out. Your scheme means they might have to drive miles on possibly blocked roads, to get to an access point that might be miles away from the obstruction.

 

[Winthorpe]

And where are the repair teams going to be based whilst the drone is flying around?
At least at the moment as soon as the proving train gets to an obstruction the team on-board get off and try and sort it out. Your scheme means they might have to drive miles on possibly blocked roads, to get to an access point that might be miles away from the obstruction.

It’s not meant to replace anything that happens already. To supplement it. Would you not prefer to know what problems lie ahead?

I made the suggestion in response to a long thread speculating on why the North Wales coast line was blocked. No one knew. We would in a lot more detail what problems there are with my suggestion.

 

[Bald Rick]

Aerial maps are produced by constant drone surveys.

I’m just suggesting it could be used to survey disrupted routes after storms.

the trouble with drones is you‘d have to have waited for the storm to pass. Flying them at low level in the tail end of the storm would be, err, ‘courageous’.

Meanwhile the route provers will go out in the storm.

 

[Llanigraham]

the trouble with drones is you‘d have to have waited for the storm to pass. Flying them at low level in the tail end of the storm would be, err, ‘courageous’.

Meanwhile the route provers will go out in the storm.

Quite.
As far as I can see the flying of drones wouldn't have helped at all.

 

[Winthorpe]

the trouble with drones is you‘d have to have waited for the storm to pass. Flying them at low level in the tail end of the storm would be, err, ‘courageous’.

Meanwhile the route provers will go out in the storm.

Not much point proving the route blocked by fallen trees while there’s gale force winds still blowing!

Surely permanent ways can be assessed by drones after the wind has died down.

Syria and the Ukraine are being bombed to buggery by these things. They must be up to the job of helping to survey disrupted rail lines.
It’s just an extra pair of eyes to work out what needs doing.

 

[cool110]

Not much point proving the route blocked by fallen trees while there’s gale force winds still blowing!

Not in the middle of the storm, but there is a sizeable window between the winds slowing enough for further tree fall being very unlikely and allowing drone flight.