New Railway Switch Developed

[whole diary]

[Lee]

Engineers at Loughborough University have developed a new railway switch which could cut failure related cancellations by as much as 90 per cent - https://www.insidermedia.com/insider/midlands/engineers-develop-railway-innovation

The new Repoint switch uses three motors to lift and shift the rails from left to right to change the direction of an oncoming train.

[eightf48544]

Interesting. not sure how adding lift into the workings improves reliability. Most point failures are lack of detection, which relies on very close tolerance to ensure that one switch rail is locked against its stock rail and the the other has a gap sufficiently wide to allow the wheel flange to pass through, and that  the FPL^▸ (Facing Point Lock) is engaged to ensure the point doesn't move.

[SandTEngineer]

It reduces friction (obviously) during the switching travel but its a sure bet something will get wedged under the switch rail once its lifted......

After nearly 50 years in the S&T^▸ (Signalling and Telegraph) industry I can tell you the best way to stop such failures is to maintain the points properly in the first place!

There's a video here showing it in operation:
https://youtu.be/kZASTbnkjgQ

Good to see some hi-tech rolling stock in that............

[Bmblbzzz]

So it avoids single point of failure by adding extra actuators, is what I gather from that video. I didn't gather how though or what lifting has to do with it (is it part of the locking mechanism?).

[stuving]

So it avoids single point of failure by adding extra actuators, is what I gather from that video. I didn't gather how though or what lifting has to do with it (is it part of the locking mechanism?).

I think the idea is that the action of the switch rail dropping all the way down serves as a lock, and its vertical position can be used for detection. That in turn is sensed at each of the three actuators, giving triple redundancy for detection as well as actuation. However, I would have thought that you could triplicate the point motors and lock detectors, and most likely the locking mechanism too if you put your mind to it.

[TonyK]

The big difference seems to be that two parts can fail and the train still runs. The lift and drop method could (and I didn't spot how it did it) reduce wear and tear. I'm of the school of thought having two engines on an aircraft merely doubles your chance of an engine failure, but let's see it follow the usual route of intense testing, then deployment somewhere quiet with intense scrutiny, then somewhere with a bit more traffic. It will all work wonderfully until the Monday morning after one is installed somewhere crucial.

But just because it's never been done that way doesn't mean we've been doing it right.

[Oxonhutch]

From the video, it appears to operate via a vertically sweeping, semi-circular arc such that there is no sliding motion as in the conventional switch blade. This will be a lot easier - and less messier - to lubricate and maintain. As long as the arc actuator does not break (nothing is ever guaranteed) when the switch is at rail level, it must be either one side or the other, held laterally by the troughs into which it drops.

Its Achilles' heal is debris - particularly ballast - getting into the trough and preventing the switch from dropping home.  All three sensors would fail to detect and the route would fail.  I have had a conventional point fail detection because of a twig or leaf mulch, plus the usual ballast. I can only see this mechanism operating in a sterile environment such as a Crossrail tunnel where everything is set in concrete - not in the Great Outdoors.

[ellendune]

From the video, it appears to operate via a vertically sweeping, semi-circular arc such that there is no sliding motion as in the conventional switch blade. This will be a lot easier - and less messier - to lubricate and maintain. As long as the arc actuator does not break (nothing is ever guaranteed) when the switch is at rail level, it must be either one side or the other, held laterally by the troughs into which it drops.

The actuator is actually 3 actuators and can operate with any one of them so I don't see this as an issue.

Its Achilles' heal is debris - particularly ballast - getting into the trough and preventing the switch from dropping home.  All three sensors would fail to detect and the route would fail.  I have had a conventional point fail detection because of a twig or leaf mulch, plus the usual ballast. I can only see this mechanism operating in a sterile environment such as a Crossrail tunnel where everything is set in concrete - not in the Great Outdoors.

I suspect this is correct.  Lets await the results of some trials. 

[Red Squirrel]

I'm of the school of thought having two engines on an aircraft merely doubles your chance of an engine failure...

I'm with Roger Bacon, late of Flight International (ah!) When someone asked him why he always insisted on travelling in 4-engined aircraft, he replied 'because there are no 5-engined aircraft'.

[Bmblbzzz]

If there are two horizontal actuators and one vertical, then surely a failure of any one actuator will still mean the switch can't switch (or at least can only move in one direction)? What am I failing to understand and how badly (there's obviously something and I suspect I'm failing it very badly indeed)?

Also, the video mentions that the design of switch used in Britain dates from before Queen Victoria's reign; this implies that other, newer, designs are used elsewhere. Newer isn't necessarily better, but why not look at those?

[stuving]

And yes, do beware those pesky jumping pebbles. 

Try this explanation:

Unlike other switching solutions, locking is provided passively by the position of the rails. The actuators lift the rails out of their lock to move them. Any single actuator is capable of providing this motion and, in the event of an actuator failure, the switch is still locked in place throughout its length.

Or else search for Repoint - but ignore all the stuff about the "stub switch layout", that's an entirely distinct concept that came out of the same project.

I'll repeat what I said before: you could put three modular actuators one a standard sliding switch, and three lock detectors, and I suspect more than one lock too. So most of the claimed advantages are not tied to the one big new thing - passive locking. And yes, do beware those pesky jumping pebbles.

[Bmblbzzz]

So they're not claiming it will still move, just that it will still be locked (and detected as locked, I presume). Understood now (I hope).

[stuving]

So they're not claiming it will still move, just that it will still be locked (and detected as locked, I presume). Understood now (I hope).

No - they are!!! There are three, identical, redundant, triplicate, actuators, each of which detects lock on both switch rails by virtue of them dropping into their locating slot.

[Bmblbzzz]

Yes, I understand the locking part. But if any one of those actuators fails, the switch will still move left, right, up and down?

[ellendune]

Yes, I understand the locking part. But if any one of those actuators fails, the switch will still move left, right, up and down?

If any two of those actuators fail the switch will still move it is using an arc so it is a rotating motion to do all the directions.